Social listening

“CoLd eMaiL tO E-cOmMeRce is DeAd!” ok. here’s how we generated 45 leads in 20 days for this ecom agency: 1. Create a STRONG frontend offer (free asset/work) Everyone is spamming e-com with the same crappy pitches. “We guarantee to increase email revenue by 50% in 90 days or u don’t pay!” “We’ll create a landing page that increases conversions by 20%!” “We’ll 2X your ROAS or I’ll buy you a steak dinner!” This slop has been pushed since 2022, THIS outreach is dead. What’s not dead is actually offering something people want, regardless how saturated the industry is. What that looks like: - Free case study masked as a “playbook” - Free sample of your product/service - Free micro service that isn’t your main service (just a lead magnet) These work, and brands still want them when positioned right. 2. Hyper-qualify lead lists E-commerce data is typically pretty bad for some reason across most databases. Even when you filter for 11+ employees, you’ll run into brands that somehow have less than a $1k/mo budget for marketing spend lol. So you need to qualify further than the base filters: - Run claygent to confirm is an Ecom brand - Score the brands online presence (signal whether have budget or not) - Filter for E-com technologies (Shopify, WooCommerce, etc) - Verify whether they’re actively running Meta , Google, LinkedIn, Apple Ads or not with Apify Apollo.io alone is not going to give you a well targeted list, need to qualify further. 3. Very short & casual messaging Compared to most industries, E-commerce is probably the most “laid back” when it comes to messaging formalities. From what we’ve seen, the more formal you are the worse emails perform. As well as email length – these guys are receiving 20-100 cold emails/day, they need to be able to understand what your entire email is about in <5 seconds or they’ll just skip it. The best performing email we’ve ever sent to E-com brands was a one-line email that is less than 20 words. It’s along the lines of “John, interested in XYZ thing I created for your brand?” Print leads and meetings, which really came down to the offer. Message just presented the offer in a quick and simple way. — Email to ecommerce is really not the beast a lot of people in the space make it out to be, there’s just less margin for error. Follow these frameworks and you’ll be golden.

I counted the platforms we touch to get one GTM outcome. Five. Claude, re-prompted because the answer wasn't quite right. Cursor, writing into repos nobody has fully mapped. Zapier, holding a workflow together that one person understands. Claygent, enriching records nobody has audited. LangChain, running agents that don't talk to each other. Each one is good at its own job. The problem lives between them. None of them share memory, so context you established in one is invisible to the next and you rebuild it by hand every time. That rebuild is unpaid work nobody puts on an invoice. Add up the subscriptions. Then add the hours spent moving context between them. That second number is the real invoice.

At our recent NYC meetup with Clay, Head of AI Jeff Barg, ML Engineer Vyshnavi Khota and Software Engineer Soroush Khadem shared how they scaled agent evals to 300 million runs a month. Watch the full talk to learn: ✅ How Claygent and Sculptor got to production scale ✅ Clay's four quadrant framework for agentic evals ✅ Why closing the loop between production and offline evals is the hardest part ✅ How a data lake and long context are changing what agents can do with data Check it out on YouTube: https://lnkd.in/grStyZ4m

Before buying another AI sales platform, choose one bottleneck and one tool. - Slow account research: test Claygent. - Weak email drafts: test Lavender. - Missed reply handling: test Make with an OpenAI classifier. - Incomplete call notes: test Fathom. - Poor prioritization: test HubSpot Lead Scoring or Common Room. Run the experiment for 30 days. Measure the baseline, introduce the tool into one defined step and review a human sample. The metric must follow the work: valid accounts per hour, positive-reply rate, speed-to-lead, missed opportunities or completed next actions. More output is a productivity gain. Revenue operations needs the downstream metric to improve.

I built a fully automated B2B outbound engine inside Clay from scratch, taking raw firmographics all the way to verified decision-makers and custom cold emails. To make the scenario realistic, I picked an existing YC-backed developer tool as my benchmark case study. (Vendo (YC S26) it is) Here's how the end-to-end system works and why I structured the scoring logic this way: → Target Definition: Sourced US-based B2B SaaS accounts (11–200 employees) experiencing feature backlog fatigue and limited engineering bandwidth. → Contact Waterfall: Built a dynamic buyer hierarchy hunting for Product Leaders (Head/VP of Product) first, falling back to Engineering Leads, and defaulting to Founders, merging them into a single clean Decision-Maker column with verified LinkedIn URLs. → Custom Scoring Engine: Designed a multi-layered formula matrix where every weight directly reflects technical fit and intent: • Size Score (+2/1): Weighted 50–200 employees higher (+2) due to immediate request volume, while 11–49 (+1) signals high growth potential. • Industry Score (+2): Prioritized B2B SaaS & DevTools where in-app customizations are a core requirement. • Core TechStack (+2): Targeted modern frontend frameworks (React, Next.js, Vue, TS) where Vendo (YC S26)'s UI embeds seamlessly. • API & Infra (+1): Verified modern backend infrastructure (GraphQL, OpenAPI, Vercel) ready to power user-built micro-apps. • Integration Score (+2): Detected active marketplaces/directories using Claygent, a direct proxy for custom request friction. • Hiring Score (+2): Scraped roles for Product Managers, Solutions/Integration Engineers to identify teams spending heavily to manage client requests manually. • Funding Score (+1): Flagged recent capital raises to identify teams with budget and active growth mandates. → ICP Tiering & Validation: Summed all weighted signals to categorize accounts into Hot (>=10), Warm (7-9), and Cold (<7) buckets, verifying emails via Enrichley + MX domain checks. → Strategic Copy Generation: Generated 2-paragraph cold emails (under 90 words, zero sales fluff, no em dashes) paired with dynamic binary CTAs, driven by separate Claygent prompts for body and subject lines. → Dual-Table Architecture: Offloaded heavy scraping and formulas into a backend Lookup Table, pushing only 4 clean fields (Domain, ICP Fit, Subject Line, Email Body) to the primary execution view. Recorded a raw Loom walking through the entire table architecture step-by-step. Link is in the comments. I'm actively taking on GTM Engineering projects. If you're looking to automate your outbound stack from the ground up, let's connect via DM or 📩 workwithdeblina@gmail.com Nour Zahzah Yousef Helal, would love to know what you think of this setup for Vendo (YC S26)! And a special shoutout to Yogesh Jaiswal for always guiding me and having my back.

The future of lead generation isn't just finding contacts. It's understanding them before you reach out. That's where AI research agents are changing B2B prospecting. In the past, the workflow looked like this: Find a company → Find a decision-maker → Get contact information → Send outreach. Today, modern AI tools can help you go further. They can assist with: • Researching companies automatically • Understanding what a business does • Identifying relevant decision-makers • Qualifying prospects based on your ICP • Finding useful signals and context • Preparing personalized information for outreach For example, tools like Clay and Claygent are making multi-step prospect research and GTM workflows more automated. But here's my opinion: AI can research faster. Human expertise still decides what matters. A tool might find hundreds of data points about a company. But an experienced lead generation specialist can ask: Does this company actually fit the client's ICP? That's the difference between collecting information and creating opportunities. The future of B2B prospecting will combine: AI Research + Quality Data + Human Strategy And I believe that's where the biggest advantage will come from. Are you using AI for prospect research yet? #LeadGeneration #B2BProspecting #AI #Clay #Claygent #SalesIntelligence #DataEnrichment #GTM #SalesDevelopment #BusinessDevelopment

Here are the 3 layers of a complete AI outbound system for 2026. It starts before the first email and keeps working after the reply. 1. Research, strategy, and data Product-market fit comes first. Then find a cold-outbound message. It may differ from what works in paid or inbound. Build the ICP from: • Sales-call recordings • Closed-won deals • Customer interviews • AI market research Sales calls carry the best evidence. Claude or Perplexity add context. Segment the ICP and map the TAM. Build account lists with Apollo, Clay, Google Maps, signals, and CRM data. Then enrich the contacts. 2. Infrastructure and multichannel outreach Set up secondary domains, accounts, two rotating batches, and a reserve pool. Accounts eventually fail. The reserve prevents downtime. Porkbun handles domains. ScaledMail handles accounts. EmailBison runs campaigns. Run three campaign types: • ICP-segment cold campaigns • Evergreen signal campaigns • Closed-lost retargeting Cold campaigns take 80-90% of the volume. Calls expose objections quickly. LinkedIn through HeyReach supports email and calling, then nurtures engaged people. 3. AI agents and RevOps Use Masterinbox to combine email and LinkedIn replies. Use Claygent or Claude Code to score replies for intent and account value. Alert sales when someone is interested. Give each account tier a different follow-up. Route every lead to the right sales team and update HubSpot or Salesforce. Then send sales-call recordings back to the first layer. Research → outreach → replies → sales calls → better research. That feedback loop turns separate tools into one outbound system. Repost ♻️ this someone who needs it P.S. Where does yours break today: before the list, during outreach, or after the reply?

In 2026, no GTM team should still be enriching data manually. Everyone obsesses over AI and automation in GTM. But none of it works without good data. The teams that: Have their whole TAM mapped and enriched in their CRM Track multiple intent signals live Identify the right decision-makers Act on those signals with the right strategy will always have an unfair advantage. What I see from a lot of GTM teams is that they still rely on outdated, legacy enrichment tools and end up overpaying for basic, unfiltered data. After testing dozens of GTM and data tools, here are 7 worth knowing: 1️⃣ Clay → 200+ data providers in one workflow → AI-powered research via Claygent → Waterfall enrichment for emails and phones 2️⃣ Freckle.io → Sits on top of your CRM and auto-enriches records → 40+ data providers + AI agents → One of the best CLIs for GTM data 3️⃣ Cargo 🧱 → Identifies and enriches key stakeholders at target accounts → Automates data flows across your revenue stack → Built for complex RevOps workflows 4️⃣ Unify → Surfaces warm leads hiding in your existing CRM → Enriches with intent + firmographic data → Connects sales, marketing and AI 5️⃣ Landbase → GTM data to power your AI → AI-automated scoring and qualification → Hiring, funding and tech-stack signals 6️⃣ Common Room → Tracks engagement across Slack, GitHub, LinkedIn + more → Enriches contacts from community and product signals → Connects intent to your CRM automatically 7️⃣ HutLaunch 🚀 → Finds and enriches your ICP across multiple channels → Monitors buying signals and intent in real time → Identifies the right people to contact → Turns signals into personalised outreach → Runs LinkedIn, email, Reddit, X and other channels together → Automates follow-ups and engagement → Turns GTM data into qualified conversations and meetings The important distinction: Most tools help you find the data. HutLaunch helps you act on it. For example: A target company: → Raises funding → Starts hiring for your target role → Changes its tech stack → A decision-maker becomes active on LinkedIn → Engages with relevant content Instead of adding that company to another spreadsheet, HutLaunch can turn those signals into an actual GTM workflow: Signal → Account → Decision-maker → Personalisation → Outreach → Follow-up → Meeting That's where GTM is heading. Not more data. Better signals + better timing + better execution.

A practical workflow for using Deepline skills to validate technographic signals before deploying scripts and Workflows.

Another successful utility locating project completed in Delray Beach, Florida! 📍 Using Ground Penetrating Radar (GPR), DeepLine FL accurately located and marked underground utilities, helping our client work safely and avoid costly damage. We find the path before you drill. 📞 (754) 272-7401 ✉️ sales@deeplinefl.com 🌐 www.deeplinefl.com #DeepLineFL #GPR #UtilityLocating #GroundPenetratingRadar #DigSafe #SouthFlorida

For junior / new GTM engineers, the market is quite interesting right now, but im seeing a problem: There has never been more demand out there, but you can no longer sell "cold email" or "outreach" as as service, nor "revops" and AI has made operations much easier & tools are cheaper than ever. So: You have to sell the outcome; and to sell the outcome though you have to understand your buyer. They're not "struggling with pipeline", they just build random lists, because they don't have their TAM mapped out. They're not "struggling with closing", they just have no system in place to derive meaning & learn from their outbound & sales calls, so their understanding of their buyer does not improve over time and the Clay hype is dying down; a GTME is not new anymore as an idea. Most companies still don't know how to scope a GTM role, but at least they know it exists. So they want it IN HOUSE. That means that while demand is going 📈 , companies are pushing hard on in house only. This means that you can either go in house, which means risking becoming too one sided or falling behind if the company is not AI native or staying in the Agency / Fractional GTM world, which means you have to become great, since good won't cut it anymore. Companies like Deepline & prospeo.io have largely democratized the infrastracture/data part of GTM & the existence of MCPs means that the "executor" GTM people / juniors are essentially useless right now, since a properly scoped task by a senior GTM person, who knows how to do strategy, needs no executor, since an agent can do all the execution & implementation. So that leaves us with 2 important trends for GTM engineers: - You arguably have to aggressively upskill to become senior enough not to get crushed by agents in 6-12 months. - You have to either go inhouse and accept the risks & benefits OR stay fractional/agency and accept that only the top half will have a chance to thrive Greetings from Athens, GR 👋

My first time running whilst hacking 🏃 I really enjoyed working with Julian Savitch-Lee and Aidar Bekchintaev to build Serenetripity at RUN/HACK London, Europe's first running hackathon. It is an AI journey planner for your mental wellbeing: you tell it how you feel and how you want to feel, it builds you a route through the city, records the walk, and learns from what you write afterwards. The rule was that you could only build the product while you were out running. So we took it in turns, dictating every change to our AI agent mid-run. I ran 15km by the end of it, made longest in a while. Three key takeaways: 1. Pushing yourself to the limit in two different capacities is positively reinforcing. 2. Tying particular actions to physical activity bouts helps prioritise tasks effectively. 3. Time spent thinking about a prompt before you run it pays for itself. We deployed more in the same time frame because of it. Try it here: serenetripity.vercel.app Thank you to the amazing organisers Tijs Nieuwboer, Luke Balabanovic, Aruzhan N., Siena Kinsale, Rachel Macnaghten 🌀, and Abdelaziz Zizou Brahmi for making this happen. Healf · Cognition × The Interaction Company of California (Poke.com) · Deepline · ROXFIT · ElevenLabs · algosoup · Thrad · Wispr Flow · Tavily · Accelerate ME · PerfectTed · Running Hackathon

64km ran around a track while sending 48 prompts from our phones to our laptops... That's what it took me, Terry Huang , and Keanu Czirjak to build our product at the world's first ever Running Hackathon. Signed up midnight the night before... showed up with minimal expectations... and ended up having a blast while running a full marathon. It absolutely poured rain… but there were saunas, pizzas, DJs, and an atmosphere that somehow stayed electric, competitive, and supportive all at the same time. Huge thank you to the organisers who pulled it off - Tijs Nieuwboer, Luke Balabanovic, Aruzhan N., Siena Kinsale, Rachel Macnaghten 🌀 and Abdelaziz Zizou Brahmi. And to the sponsors who kept everyone fed, caffeinated, building, and moving: Healf, Cognition, The Interaction Company of California (Poke.com), Deepline, ROXFIT, ElevenLabs, algosoup, Thrad, Wispr Flow, Tavily, Accelerate ME and PerfectTed. Was also great meeting and competing with so many people: Samuel Zhang, Javiera Rubio, Andrey Lebedev, Arlind V, Kyna Jain, firdavs olimjanov, Umair Husain Khan, Kaushik .K, Fergus McKenzie-Wilson, Siddarth Oruganti, Yasser Boughou, Adithya Raghuraman and so so many more!

The real lesson in Garrett Wolfe's April Tools Day talk is that the signal and the action should live in the same loop. See how Deepline helps GTM teams run

Claude Code prompt: Build 5k.vc: the free, no-signup investor graph covering the 5,000 VC firms that matter, their actual investing partners, portfolio companies, investments, rounds, co-investors, sectors, geographies, stages, check sizes and outcomes. Starting with zero data, create an autonomous public-web ingestion/enrichment pipeline that discovers and continuously improves this graph; obsess over canonical entity resolution, deduplication, investment dates/rounds/amounts/lead/partner attribution, provenance and confidence - accuracy matters more than raw record count. Build beautiful, extremely fast, SEO-indexable pages for every firm, investor and funded company, plus searchable/filterable directories, rankings, relationship graphs and “what they actually do” analytics based on observed investments rather than stated thesis. Make the homepage product “Who should fund you?”: accept any company URL/name/description and rank the investors and specific partners most likely to invest based on stage, raise size, geography, sector, recency, check size, lead behavior, relevant past deals, co-investor network and portfolio conflicts, explaining both why and why not for every result. You have full freedom on architecture, stack, data sources, algorithms and design; build the real working product end-to-end rather than a mockup, seed it with as much high-confidence public data as you can, keep it 100% free/private/no tracking, and keep iterating autonomously until it feels like the definitive investor intelligence product founders would use instead of manually researching VCs. Match it with founders.io.

🇺🇸 Here’s how I’d actually build a lead generation system with Claude. I wouldn’t start by adding every tool I can find. I’d start by asking one question: What parts of lead generation do I want Claude to handle for me? For most teams, that usually means research, enrichment, prioritization, outreach, and follow-up. That’s where this stack starts to make sense. Plugins can help with prospect research, enriched lead lists, messaging, and marketing data. Skills can make those tasks repeatable: → Turn an ICP into a prospect list → Research target accounts → Enrich lead information → Draft personalized outreach → Prepare for sales calls → Prioritize warmer opportunities → Build cold email sequences Then MCP connections bring everything together. Claude can connect with tools like Apollo, HubSpot, Clay, Common Room, Notion, Slack, and Zapier to pull context and move work across the GTM stack. So instead of doing this manually: Research → copy data → enrich → check CRM → write message → update pipeline You can build a much more connected workflow: Research → Enrich → Prioritize → Personalize → Outreach → Track That’s the part I find interesting. Claude becomes far more useful when it has the right tools, context, and workflows around it. The goal isn’t to install 21 things. The goal is to build a system that removes the repetitive parts of lead generation while keeping the important decisions with you. Which part of your lead generation workflow would you automate first?

Companies bought Copilot to keep files inside Microsoft 365. Staff opened Claude anyway. Microsoft said Microsoft 365 Copilot passed 30 million paid seats. That is a real number. Security teams could say yes because the questions stayed in the company Microsoft account. The product still did not feel like ChatGPT. People opened Claude. They also kept a personal ChatGPT window. Reco found four in five AI tools in the companies it watches had no IT approval. Ramp’s May index is the punchline. Among the businesses Ramp tracks, 34.4 percent pay Anthropic and 32.3 percent pay OpenAI. Those two shares add to more than 100 percent. A lot of those firms write checks to both. Salesforce, ServiceNow, and Amazon Quick now let you pick the same models. Amazon Quick is the standalone work assistant. It connects to Microsoft 365, Google Workspace, Slack, and Salesforce. You do not have to move the work onto AWS first. Nova is a different product. That is Amazon’s own model family on Bedrock. Grok is fast when I use it. Getting it on an approved-vendor list is the hard part. I used machine learning to score leads years before ChatGPT. We used a model on deals we had already won, then scored new leads against that list. Writing an email is a different job from guessing who will buy next. Vendors sell all three as “AI.” Question for CIOs: list the models your official software already uses. Then list the models your staff open on their own. How big is that gap? Full piece: https://lnkd.in/e5eYQc_j #EnterpriseAI #Copilot #Claude #Microsoft365 #CIO

Si te preguntan por tu RAT, tu canal ARCOP y tu portal del titular... y tienes que abrir 3 Excels, un Drive y el WhatsApp del DPO... No tay tan securo. En Chile la Ley 21.719 no te pide un brochure de privacidad. Te pide operación: - Registro de tratamientos con juicio firmado - Derechos ARCOP con plazos y evidencia - Portal público para el titular, en tu dominio - Sala lista para cuando te fiscalicen Eso es Securo. Cumplimiento inteligente. El sistema donde dejas de improvisar el cumplimiento y lo operas en un solo lugar: ARCOP, incidentes, encargados, transferencias, EIPD, comité, retención y pack de trabajo. ¿Y por qué "inteligente"? Securo tiene servidor MCP. Lo conectas a tu Claude Code, Cursor, Codex o al IDE que uses, y se pone a armar tus RAT, inventario de datos y mapa de tratamientos directo desde tu repo. Sin inventar. Lee tu código, detecta qué datos tratas, dónde y cómo, y deja el borrador listo para que tu DPO firme el juicio. Pasas de documentar a mano a tener evidencia viva. Securo no es un sello ni una asesoría legal. Es tu sistema de trabajo. Donde queda el expediente vivo: qué tratas, con qué base, quién lo revisó y dejó constancia. Listo para operar y para mostrar cuando te pregunten. Si eres DPO, admin o tech lead y te toca implementar la 21.719 de verdad, conversemos. securo.cl

All roads lead to metered pricing. 📉 The era of flat-rate AI is over. 🛑 Cursor notified users this week about their new billing structure. Auto pricing will now charge based on the specific model each request uses. 💸 This is part of a massive industry shift. GitHub Copilot is retiring its annual plans. 🗓️ They transitioned to token billing in June 2026. Every input, output, and cached token is now drawn at API rates. 📊 Anthropic is restricting rather than repricing. Weekly caps arrived on Claude Code back in August 2025. ⏳ Enterprise users pay a seat price plus usage metered from the very first token. Enterprises are feeling this collision right now. 🏢 Procurement runs on strict annual cycles. Pricing models are changing quarterly. 🔄 A budget approved in Q1 2026 is already completely obsolete by June. Everyone is metered already to some degree. The critical question is what it will cost you to leave. 🤔 Can you point your agent at a different model tomorrow without a massive surcharge? 🔌 If you bring your own key, are you still paying the vendor per token? Is your agent an extension or a full editor fork that forces your whole team to switch IDEs? 💻 Do your prompt libraries and configs live in your own repo or a vendor marketplace? What actually happens to your prepaid credit balance if you walk away? 💳 Vendor lock-in is getting incredibly expensive. You need to start building flexibility into your team workflows today. 🛠️ Are you prepared for the reality of metered AI costs? Join our Practical AI group on Skool to dive deeper into these strategies and connect with other builders. 🧠 ♻️ Repost this to help your network rethink their enterprise tool budgets. ➕ Follow Deven Goratela [https://lnkd.in/dQwsb2jA) for the latest insights on staying ahead in AI and automation. #ArtificialIntelligence #SoftwareEngineering #TechTrends #Copilot #Cursor #Anthropic #MeteredPricing #PracticalAI

It's crazy how many people still run their LinkedIn accounts fully manually. Here are 11 things to delete and the 11 Claude skills that fix it: (all 11 skills are free, check the image) Delete these from your profile today- 1. Coment CTA's in the post copy 2. Hashtags. Every one 3. Stacked asks. Like, comment, connect and repost 4. Links in the first comment. Put them in the post 5. Long hooks. Both lines have to be seen on mobile. 6. The pitch in the first message. 7. Value / Info posts. 8. "Helping founders scale" in your headline. 9. Only focusing on cold outreach. 10. Waiting for people to message you first. 11. Posting five days a week with nothing to give away These are my personally used 11 Claude skills that fix these: 1. /icp Builds the buyer profile the whole system runs on. 2. /profile Rewrites the headline, banner and about section as a landing page. 3. /requests Writes the connection request. The accepted request is the distribution. 4. /ideas Turns this week's live work into post ideas across all 27 structures. 5. /hooks Writes the first two lines so both land before the mobile cut. 6. /post Writes the full post in your voice, not a version of your voice. 7. /audit Paste a draft. It comes back rewritten, not scored. 8. /lead-magnet Builds the resource and the image the CTA lives on. 9. /comment-to-dm Drafts the replies to everyone who commented, one ask each. 10. /outreach 53 messages across 11 stages. Connection request to booked call. 11. /numbers Content leads and outreach leads counted separately, every Friday. I'm giving away the complete library for free. https://lnkd.in/edcu2yka

🎯 Weekend Learning & Course Completion Glad to share that I’ve completed Building with the Claude API by Anthropic! 📚 This was quite a comprehensive course covering multiple small but important concepts around building and working effectively with the Claude API. Each concept added something new to my understanding of how AI applications can be made more efficient and effective. 📌 One interesting concept I learned was Prompt Caching: Imagine repeatedly sending the same large document to Claude while asking different questions. Processing the same content again and again can increase both response time and cost. Some ways to make this more efficient include: • Prompt Caching – Reuse previously processed context to reduce repeated processing and improve efficiency. • Cache Breakpoints – Define where reusable content ends and new content begins for more effective caching. • Splitting Content – Break large documents into smaller, relevant pieces when appropriate. • Asking Multiple Questions Together – Combine related questions to reduce repeated interactions with the same context. 💡 Key Takeaway: Building with AI is not just about sending prompts and getting responses. Understanding concepts like prompt caching, context management, and efficient API usage can help build faster and more cost-effective AI applications. It was a big course with many small concepts, but every concept contributed to a better understanding of how to work with AI APIs effectively. You can also spend a few hours over the weekend learning a new skill-small and consistent efforts can lead to meaningful growth over time. #WeekendLearning #CourseCompletion #ClaudeAPI #Anthropic #AI #GenerativeAI #PromptCaching #AIEngineering #Upskilling #ContinuousLearning #LearningNeverStops

📊 Continuing my Power BI learning journey! I asked Claude AI to generate a practice dataset for me a star-schema Electronics Retail sales dataset with fact and dimension tables so I could keep sharpening my skills beyond tutorials. Here's a breakdown of what the analysis shows: 💰 Revenue: ₦1.33bn | Total Cost: ₦1.03bn | Profit: ₦291.98M across 250 customers 📉 Revenue by Year: Dropped from ₦683M (2023) to ₦642M (2024) — a trend worth digging into further. 🏆 Profit by Brand: Apple leads with ₦82M in profit, followed by Asus (₦75M) and Samsung (₦47M). Lenovo and Dell are tied at ₦31M, with HP trailing at ₦27M. 🎨 Total Cost by Color: Black-colored products account for the highest total cost (₦0.56bn), more than double Gray (₦0.29bn), with White and Silver much lower. 📆 Customer Activity by Month: Customer count fluctuates between 56 and 78 monthly, with December and October showing strong peaks. 🍩 Profit by Income Level: Medium-income customers drive almost half of all profit (48.28%), followed by Low income (28.31%) and High income (23.41%). 🗺️ Revenue by Region: South leads (₦0.35bn), closely followed by North and West, with Central trailing at ₦0.15bn. Grateful to TS Academy and Ezekiel Aleke for the guidance on this learning journey.

GrowSEOService – Helping Businesses Get Found Everywhere. Your customers are searching every day—not just on Google, but also on ChatGPT, Google AI Overviews, Gemini, Perplexity, Claude, and other AI-powered search platforms. At GrowSEOService, we help businesses increase online visibility with future-ready SEO strategies. Our services include Technical SEO, Local SEO, On-Page SEO, Off-Page SEO, E-Commerce SEO, GEO (Generative Engine Optimization), AEO (Answer Engine Optimization), Entity SEO, Semantic SEO, and AI Search Visibility Optimization. We don't believe in shortcuts. We use ethical, white-hat SEO strategies that build long-term rankings, qualified traffic, and sustainable business growth. Whether you're a local business, eCommerce store, startup, or enterprise, we create customized SEO strategies designed to improve your search presence, generate more leads, and increase conversions. If you're ready to grow your business and become visible across both traditional search engines and AI search platforms, GrowSEOService is here to help. Let's build your online authority and grow your business together. Rahatul Islam SEO Expert | GEO • AEO • LLM • Advanced SEO #SEO #GEOSEO #AEOSEO #LLMSEO #AISEO #SemanticSEO #EntitySEO #localseoservices #seotipsandtricks #seostrategy2026 #seoservices #seoexpert #SEOSuccess #SEOSpecialist #seoconsultant #seostrategies #growseoservice #SeoServiceProvider #seoexpertinbangladesh

Most sponsors, deal teams, and family offices trying to prompt Claude or ChatGPT for LP research run into the exact same wall. AI is a natural cheerleader. If you ask, "Is X an investor in my space?" it hallucinatingly says yes to almost everyone. To turn LLMs into actual allocator research analysts, you cannot ask them to find matches. You have to build an Adversarial Architecture—forcing the model to act as a harsh, multi-stage filter. We did not just build this in theory. Over the last few years, we aggregated validated mandate data across real asset deployments to build (investwithsumair.com), proving this exact framework in live capital markets. We are now expanding this engine into VC, M&A, and Private Equity. Here is the 7-layer architecture required to turn raw prompts into a precision fundraising tool: 1. Mandate & Size Constraints Hard-coding strict parameters (e.g., Fund Size $\times$ 5–15% concentration limits) cross-referenced with recent transactions to eliminate 80% of bad leads upfront. 2. Capital Stack Stratification Classifying counterparties across 7 exact capital types (Senior, Stretch Senior, Mezz, Pref, JV/LP Equity, Co-GP, GP Stakes), screened by active investment periods, not marketing collateral. 3. Lookalike Mandate Matching Structuring prompts to evaluate 7 strict comparability rules and 22 legal entity naming patterns against verified transaction histories. 4. Blind Counterparty Discovery Targeted query angles that locate non-obvious family offices and private wealth without relying on surface-level name recognition. 5. Adversarial Verification Engine Prompting models with 12 explicit failure checks across 9 pipeline stages to force one of four precise outputs: Keep, Reject, Unclear, Unchecked. 6. Outbound Decision Matrix Classifying outreach into 7 tiers ranked by recipient friction, backed by strict regulatory boundary checks. 7. Primary Data Grounding Integrating verified primary sources- like Companies House charges registers and public pension board packets, for indisputable proof of active capital deployment. If your team is trying to build an internal AI fundraising stack and you want the exact logic parameters, prompts, and breakdown behind these 7 layers: Comment "BUILD" below or send me a 1-line DM with your current mandate (Asset/Business, Layer, Size, Geography). I will share the complete architectural breakdown and run your deal parameters through our live filter. (Allocating time for 5 mandate reviews this week). Out of curiosity - which layer of the capital stack is the hardest to source for right now? Drop it in the comments.

Claude fica muito mais poderoso quando deixa de ser apenas uma janela de chat e passa a operar conectado ao seu sistema comercial. 🟣✔ PLUGINS Revenoid: pesquisa e engajamento. Vibe Prospecting: listas enriquecidas. Apollo: prospecção e enriquecimento. BetterCallClaudeGrowth: estratégia de GTM. Claude Marketing: fluxos de marketing. Octave: inteligência de mensagens. Windsor.ai: dados de marketing. 🟣✔ SKILLS prospect: criar ICP e listas. enrich-lead: enriquecer dados. account-research: pesquisar contas. draft-outreach: personalizar mensagens. call-prep: preparar calls. lead-triage: priorizar leads. cold-email: criar sequências. 🟣✔ MCP Apollo, HubSpot, Clay, Common Room, Notion, Slack e Zapier para conectar dados, CRM, sinais, conhecimento, comunicação e automações. Como montar: 1. Defina o processo comercial. 2. Escolha só as integrações necessárias. 3. Conecte fontes de dados. 4. Crie skills por etapa. 5. Adicione CRM e automação. 6. Teste com 10 leads antes de escalar. 7. Revise permissões e dados sensíveis. PROMPT PARA USAR: “Projete uma stack no Claude para geração de leads. Meu processo é [PROCESSO]. Organize plugins, skills e MCP por pesquisa, enriquecimento, priorização, outreach, CRM, follow-up e análise. Inclua permissões mínimas e pontos de revisão humana.” #claude #mcp #skills #plugins #geracaodeleads #automacao

🚨 LinkedIn outbound just became a prompt. No lead lists. No manual research. No enrichment workflow. No writing DMs one by one. Just tell Claude: “Find 50 decision-makers at B2B SaaS companies who interacted with my competitors this week and start outreach.” Then watch it work. Claude finds the prospects. Filters them against your ICP. Researches their company + recent activity. Finds their contact information. Ranks them by buying intent. Writes a personalized message for each one. Starts the outreach. And when they reply? Claude handles the conversation, qualifies the prospect, follows up, and books the meeting. You can literally come back the next day and ask: “Who replied?” “Who should I follow up with?” “Which prospects are converting best?” This used to require 4 different tools + hours of manual work. Now it starts with one prompt. One prompt → pipeline. This is getting ridiculous. Free access below 👇

Anthropic posted about “one memory” across Chat and Cowork. Since I move context between Claude sessions and services daily — the news sounded like a lifeline. It wasn’t, not quite. In the news there’s a clean headline about it: “it works if Cowork is in the cloud.” But I kept hitting the same problem: I’d save something in a project chat, and in Cowork it was as if I’d never said it. Trying to understand the system from posts and discussions with Claude didn’t help. So I had to test it manually — using “code words” as markers. E.g. inside a project in Chat mode I saved “purple giraffe-42.” Asked in Chat mode in a new session — it came back instantly. Requested the code word outside the project — nothing. Desktop Cowork, regular “Cloud” project (no local folder attached) — nothing either. Looks like on the desktop app, even in a “cloud” project, Cowork sessions always stay “local.” Tried in the browser: same project, now cloud Cowork mode. It honestly searched memory and never found the giraffe. When I asked it to read every memory file, Cowork found 11 of them — general memories (about a car, my daughter, etc.), not info about this project. As I see it, “Claude memory” is at least two things. “Global” (what shows in Settings → Memory): Chat and cloud Cowork both read and write it — tested. And “Project memory” in Chat mode: only Chat can read and write it. Cowork never does — not even in the cloud. The cloud caveat in the posts is honest, but it isn’t enough: in reality the cloud opens only global memory, not project memory — so this memory holds everything about you, not only what’s related to this exact project. And a “local project” and a “local session” are not the same thing. Cloud Cowork only worked for me in the browser — every Cowork session in the desktop app automatically became “local,” so that shared-memory path wasn’t there. And if I start a local project with a folder — Chat mode is disabled = no memory can be used, only files on disk. A few hints: - Need a fact in both Chat and Cowork inside one project (like common project-related context)? Put it in Project Knowledge. - “Global” memory is the only shared channel, but it has no walls — stuff from different projects can pile into one heap. - Local Cowork cannot reach cloud memory at all 😔 — only files on disk unlock sharing context between sessions. “One memory across chat and Cowork” isn’t complete bs, but it can lead to confusion and wasted time and tokens. Read only the headline and you’ll think it means all memory, including project memory. That’s what I thought, until I checked. Full write-up of what I checked, plus insights: https://lnkd.in/dH7Cti82. Next I want to see how this memory lines up with Claude Code, Claude Design, and the rest 🧐

I built a fully functional CRM for my company without writing a single line of code myself. No SaaS subscription. No dev agency. No months of back and forth. Just Claude, Google Sheets, and a clear idea of what I needed. Here’s what we built for Planet Stays — our premium holiday homes business across Uttarakhand and Himachal Pradesh: ✔️ A 35-column lead and booking management system handling 10,000+ records ✔️ A custom web app UI for our ops team — search, add, edit leads with auto-generated IDs ✔️ An email parser that reads Gmail every 15 minutes and auto-creates booking entries from Airbnb, MakeMyTrip, Goibibo, and Agoda emails ✔️ Automatic cancellation and review tracking across all OTAs ✔️ 4 live dashboards: Review Performance, Cash Payments, Repeat Guests, Revenue — all refreshing daily The whole thing runs on Google Sheets + Apps Script. Cost: essentially zero. Compare this to what a proper CRM SaaS would cost: → HubSpot, Salesforce, or a custom-built solution = ₹50,000–₹3,00,000+ to set up, plus monthly subscriptions → We built something more tailored to our exact workflow, for a fraction of that The shift that’s happening is real. AI isn’t just helping developers code faster — it’s letting founders and operators build things themselves that previously required a full tech team. I didn’t need to know JavaScript. I needed to know my business deeply enough to describe exactly what I wanted. Claude handled the rest — including debugging, edge cases, OTA email format variations across 4 platforms, and 4 custom reports. A few things I learned: 1. The more precisely you can describe your problem, the better the output 2. Iteration is fast — what would take a dev team weeks took days 3. You end up understanding your own system deeply because you’re involved in every decision 4. The maintenance SOP is a Word doc — also written by Claude SaaS isn’t dead. But for specific, well-defined internal tools? The calculus has changed. #AI #Productivity #Startups #NoCode #Claude #HospitalityTech #Founders

Happy Sunday corporate goyim slaves. It is i, the black panther of Wakanda Chase; with my teacup golden doodle companion Kobe. 🌴⚡️👨🏽‍💻🦁 I suppose i am retired now based on the fact the crooked mother fuckers The Wawanesa Life Insurance Company and their lawyer Steven Abramson of Harper grey who I caught lying in writing (council corroboration) on the record; will be sending me at least 3 million dollars or I will take them to the judge. Therefor I am no longer psychologically caged to behave like a good goyim in the corporate world of LinkedIn. I have shipped upgrades to NomadTechnologist.com I will be covering all of the policy changes that the covert virus like Zionists that occupy all of our governments are using against us. While their politicians puppets ot they themselves lie to our faces. If there is any topic you would like to cover. Like the holocaust genocide in gaza. The poison in our foods. The rapists that use Roblox to find children. The manager that lies in their write up to HR to cover their asses instead of taking accountability for their incompetence (true story when I worked at Boomi this happened to me hence why Colin Fenn was fired); or any other topic that a typical corporate goyim slave would never write about on LinkedIn. Look no further than here for a safe place for discussion. Corporate goyim unite. Free the white man from the attack on his race from the covert zionists who are doing their hardest to spark a race war like Jake Lang. Free the black man (and woman) from the weak humans that use abuse of positions of power to take out their frustrations of being inferior on races of colour that are actually not the reason for any of their problems. For i am a free man now. Follow the words of the NomadSignal.ai newsletter I shall lead you to the promise land. Tech goyim unite. Allegedly. 🛸 🙃 Wheres my money The Wawanesa Life Insurance Company you mother fuckers. The black panther of Wakanda will ensure a precedent if we go to court. I promise you that. Pull up. My litigation engine is serious. Ask Anthropic they’ve been trying to train Claude code on my prompting to improve their product. Ask Palantir Technologies they were spying on me on behalf of Boomi Francisco Partners Operating Executives Mark Zuckerberg Meta YouTube and others. Don’t think I forgot about the racist trolling email with the customer service agent named Jacobi. When I screenshot it and sent it to my lawyers in email. Then the email agent name was changed. Emails from Replit in mt inbox. The fucking customer service rep name was Jacobi. I screenshot it and sent it to my lawyers. Then it changed to a generic name. I still have the screenshots. Anyone ever had an old email already sitting in your inbox have an update where the from email name was changed? Magical. Goyim unite.

AI isn’t just about asking questions anymore — it’s becoming a skill that can be applied across learning, creativity, analysis, and career development. I recently had the opportunity to attend a two-day practical training at the ASU Career Center focused on Artificial Intelligence and its real-world applications. Day 1 | Exploring AI & Building with Prompts The first day was all about understanding how to work effectively with AI. We explored different AI tools, including Chat GPT, Claude, Notion, Manus, Canva Ai and Gemini, and learned how better prompts can lead to better results by clearly defining: • The Role • The Task • The Context • The Expected Output Instead of stopping at theory, we put these concepts into practice. One of the projects was creating a webpage concept for “X,” a men’s Fashion brand, using AI to support the design and development process. We also explored how AI can be used in Data Analysis and Statistics, which showed me how useful AI can be when working with data and turning information into meaningful insights. Day 2 | Turning AI into a Productivity Tool The second day focused more on using AI as part of our everyday workflow. We explored different ways to use AI for: • Organizing studies and projects through Notion • Creating presentations using Canva Ai • Use Higgsfield AI to make video content • Improving and tailoring CVs based on specific Job Descriptions • Supporting different stages of a project • Making learning and research more efficient What stood out to me throughout the two days was that AI itself isn’t the main advantage. The real skill is knowing what to ask, how to provide the right context, how to evaluate the response, and most importantly, how to turn AI-generated output into something practical and valuable. A big thanks to the ASU Career Center and the trainers Mostafa Hagras and Ahmed Dahy for the knowledge, practical activities, and hands-on experience. Excited to keep learning and exploring how AI can become a bigger part of the way I work and develop professionally. #ASU #asucareercenter #AinShamsUniversity #ArtificialIntelligence #AI #PromptEngineering #GenerativeAI #DataAnalysis #CareerDevelopment #DigitalSkills #FutureOfWork #AIforWork #AIforLearning #Productivity #ChatGPT #ClaudeAI #GeminiAI #Notion #Canva #Higgsfield #CareerGrowth #ContinuousLearning #ProfessionalDevelopment #Technology #AIFuture

To close the weekend, I am glad to share that I have cleared my 1st Claude certification to follow the 3 SAP certifications done in the last few weeks. And to be able to clear it with a score of 983 was the icing on the cake!! This one tests you and gets you thinking on how you can leverage Claude and AI functionally + responsibly with the right prompting, model choice, grounding techniques and validation framework. I always enjoy exploring new learning pathways, so looking forward to see where this one leads to 🙏 #ClaudeCertified Anthropic https://lnkd.in/dTazefWR

Every marketing lead we get is a demo request. Not a content download. Not a webinar list. Someone came to our site and asked to see the product. Same lead, same intent, every AE. So why does one rep turn those into pipeline while another watches them die? For most of my career the answer to that was a shrug. Territory. Timing. "The leads were softer that month." I built a Claude Artifact this quarter that made the shrug impossible. Leads in, SQOs out, by rep. The spread wasn't subtle, and it wasn't random. The same names sat at the top and the bottom month after month. So instead of guessing, I had Claude Anthropic read every intro and discovery call behind that board. Not a sample. All of them. Then compare the reps at the top against the reps at the bottom. One guardrail first, because conversion rate on its own lies. A high number can just mean a rep who qualifies loose. So we also pulled how long each SQO survived after it converted. If a rep's SQOs die three weeks later, that wasn't qualification. That was a wave-through. The top converters' deals held. The rate was real. Here's what separated them. The low converters ran clean calls. Buyer names a problem. Rep confirms it, maps our product to it, sets the next step. Textbook. Nothing you'd flag in a 1:1 (we actually have a Gong scorecard for these calls and the AE's would get 4/5 or 5/5). The top converters never stopped there. The buyer's first problem wasn't the destination. It was the door. They kept going. What happens upstream of that? Who else touches it? What breaks when volume doubles? By the end of the call there weren't one or two places we helped. There were five. Which is why the price conversation goes differently for them. Same list price. Same deck. A buyer weighing our number against one problem is doing math that usually doesn't work. A buyer weighing it against five isn't doing that math at all. And that's where the argument ends, because it's the same two names all the way down the board. Highest inbound conversion. Top revenue to date. The smallest discounts on the team. You can talk yourself into believing a high conversion rate is just a rep with a low bar. You can't tell that story about the rep who's first in revenue and last in discounting. Claude can tell me which rep stopped at the first answer. It can't ask the fourth question. That moment, where the buyer has answered you, the call feels finished, and you choose to stay in it anyway, is still the rep. Staying in a conversation that's already complete is the hardest thing in discovery, and it's most of the job. Everyone got the same hand raised. The leads were never the problem. Don't stop at the first problem. That's not the deal. That's the door. #GTM #SalesLeadership #Discovery #AIinSales

MIDDAY GLOBAL AI ANALYSIS REPORT By Ed Hernandez | KEYSTONE PULSE MEDIA / Global Media Center Sunday, August 30, 2026 | Executive Midday Pulse Edition: Sunday Digest Good afternoon. Sunday is for sorting the week, not inventing a new one. U.S. markets are closed. NVIDIA last finished Friday at $217.55, down about 4.6 percent after Thursday’s 8.7 percent jump. The earnings boom is still intact. The new questions for Monday are financing, memory, and who is allowed to use whose models. The weekend’s main product fight has not cooled. OpenAI will stop supplying models to Cursor on November 12 after SpaceX completed its $60 billion purchase. Cursor says OpenAI is only a small slice of its traffic. Anthropic is doing the opposite and adding Claude compute inside Cursor. That makes Grok 4.6 and Claude more important to the SpaceX coding stack. Anthropic still leads the model board with Mythos 5, Fable 5, and Opus 5. It also has a court win against the Pentagon blacklist. The legal picture is not clean, though. Sony Music and Warner Chappell have sued Anthropic over training data. Tomorrow matters on price as well: the Claude Sonnet 5 promotional rate of $2/$10 is scheduled to end August 31 and return to $3/$15. OpenAI’s GPT-5.6 Sol remains number four. Google is holding the workhorse lane with Gemini 3.7 Flash and Omni 1.1 tools. Meta has pulled back an internal plan to replace large parts of teams with AI agents after those agents caused disruptions. Mistral remains the sovereign-cloud path. Asia used the weekend to ship cheap capability. Tencent’s Hy4 Preview is out under Apache 2.0 at about 770 billion parameters with 49 billion active. Qwen3.8-Flash and GLM-5.3-Flash still set the low-cost floor. Kimi K3 leads Chinese-lab rankings. DeepSeek V4-Pro and Baidu’s ERNIE family hold cost and domestic-silicon share. On chips: Lambda raised $1 billion in debt to buy more NVIDIA GPUs. AMD’s Helios racks are still ramping. Huawei’s Ascend line is sold out in China and is bidding for projects abroad. SK Hynix says the memory shortage can last through 2030. The Sunday read: demand survived the week. Permission and price are what Monday will test. If GPT leaves Cursor and Sonnet costs more on Tuesday, where do your developers go? Sources Reuters, OpenAI, TechCrunch, AIdapted, BenchLM, Digitimes, company pricing pages as of midday August 30, 2026. #AI #MiddayAI #Cursor #Claude #Sonnet5 #Hy4 #NVIDIA #Qwen #KeystonePulse

If you're building a new system with both frontend and backend, and tokens aren't an issue, use team-based agents like Claude offers. Assign 3 teams: one main agent to lead and communicate like a project manager, one for frontend, and one for backend. Claude handles this well—ChatGPT can too, but I prefer Claude from experience. The main agent assigns tasks and coordinates. If backend needs a page, main contacts frontend. If frontend needs an API, main coordinates with backend. This makes development fast. You can add more agents later for code quality if needed. Set clear goals first. After implementation, check if goals are met. Then assign 2 new agents—one for code review, one for security review. Using multiple agent teams saves time. Process: write implementation plan in markdown, assign teams, set goals, implement, verify goals, then final code and security review. My advice: read Claude and ChatGPT developer docs to find even better options.

⭐ The Best Summer Of My College Journey ⭐ This summer, I had a transformative experience as a Transaction Advisory Intern at Midwest CPA. Over the 11 week internship, I immersed myself in 35 fast-paced Quality of Earnings projects, which significantly sharpened my financial analysis, reporting and communication skills. A highlight of my internship was the incredible opportunity to explore the intersection of AI and Accounting. By using Claude Cowork, I built a Claude Skill to automate Phase 1 of the QoE process: populating data into Excel templates and flagging anomalies. The Skill saved our team approximately 3 to 4 hours per project. Beyond the technical efficiency, this project taught me how to critically evaluate AI outputs. As an aspiring professional, I learned that while AI is a great tool to boost productivity, the users must be able to critique the outcomes and lead the conversation. I can only effectively train Claude when I thoroughly understand and can perform the process independently. However, the most memorable part of my time at Midwest CPA was the supportive culture and close-knit team. Beyond daily work, we built personal connections by sharing stories about our lives outside of work (Fun fact: I was inspired to sign up for my first half marathon by my incredible superwoman supervisor Alyssa Harp, CPA, MBA.) I'm deeply grateful for the exceptional mentorship and guidance from Alyssa Harp, CPA, MBA, Chris Barrett, CPA, Naiza Cuasito, and Kirt Russelle Peconada. Their willingness to answer any of my questions, guide me step by step and their professional practice accelerated my learning curve and sharpened my skills. As I step into my senior year, I will carry forward the mindset of critical thinking, a human-centered approach, and sharp attention to detail. I look forward to applying skills I learnt to my upcoming Tax, Audit, and Senior Project coursework while continuing to build meaningful connections with my professors, friends and college staff to make my last year at Luther most memorable and impactful. #TransactionAdvisory #QualityOfEarnings #QoE #AI #Claude #Internship

Free Traffic vs. Paid Traffic for Local Business – The Truth Should you keep spending money on Google Ads, or should you build a system that generates leads long after your marketing is published? In this video, I break down the real difference between free (organic) traffic and paid traffic, and explain why smart local businesses are investing in YouTube, Facebook, Instagram, TikTok, and AI search platforms like ChatGPT, Claude, and Google's AI Overviews to create a long-term competitive advantage. You'll learn: Why paid ads stop the moment you stop paying How organic traffic compounds over time Why video is becoming one of the most powerful local SEO assets How AI search is changing the way customers find local businesses How to build a predictable stream of leads without relying solely on advertising 📺 View the full video showing how we're dominating local search and AI for our local business clients: https://lnkd.in/dWn8JGFe 📅 Schedule a free strategy call: https://lnkd.in/dBwg9NUp If you want your business to be found on Google, YouTube, Facebook, Instagram, TikTok, and AI search—not just paid ads—let's talk. 👍 Like, Subscribe, and Share this video with another local business owner who wants more leads. #Hashtags #LocalBusiness #LocalSEO #GoogleAds #FreeTraffic #PaidTraffic #VideoMarketing #YouTubeMarketing #AISearch #ChatGPT #GoogleBusinessProfile #LeadGeneration #BusinessGrowth #SmallBusinessMarketing #DigitalMarketing #MarketingStrategy #TopRankedVideo

#Automating Signal-Over-Noise: How I Built a Real-Time Tech & Security Intel Pipeline Keeping up with breaking tech trends, vulnerabilities, and industry news across Hacker News, Reddit, and RSS feeds often feels like drinking from a firehose. I wanted a clean, automated solution that aggregates raw feeds, extracts high-value intelligence, and delivers concise summaries directly to my phone. Here is how I built and deployed CyberPodda SL's Automated Intel Pipeline: ##Why I Built This +The Problem: Manual feed checking consumes hours daily and introduces cognitive overload from duplicate or low-priority posts. +The Solution: An end-to-end automated workflow that fetches data, runs it through an AI summarization model (Filter → Analyze → Prioritize → Summarize), and pushes instant alerts to a custom Telegram bot. ##Accelerating Development with Claude Code Building complex automation logic manually can lead to tedious debugging. By utilizing Claude Code directly in the terminal, I was able to: +Rapidly design and refine node parameters and custom JavaScript expressions inside n8n. +Auto-generate deployment scripts and streamline Docker container networking. +Rapidly troubleshoot environment configuration edge cases during setup. ##Deployment & Infrastructure Stack To keep the pipeline robust, secure, and cost-efficient: +Cloud Infrastructure: Oracle Cloud Infrastructure (OCI) Compute Instance running Ubuntu. +Automation Engine: Self-hosted n8n on Docker with a PostgreSQL database backend. +Reverse Proxy & Security: Nginx with Let's Encrypt SSL/TLS for secure HTTPS webhook transport. +DNS Routing: DuckDNS for dynamic domain mapping to the cloud instance. +Notifications: Real-time push payloads routed directly to Telegram. Building this pipeline demonstrated how combining cloud infrastructure, open-source automation tools, and agentic coding helpers like Claude Code drastically reduces development time—turning a multi-day build into a seamless afternoon project. How are you leveraging AI and automation to streamline your daily info workflows? Let me know in the comments! #n8n #Automation #CyberSecurity #ClaudeCode #AI #DevOps #OracleCloud #Docker #BuildInPublic #OpenSource

Stop scrolling—your next massive career upgrade is officially here. Top global enterprises are scaling aggressively right now, and we have opened high-impact roles across Cloud, AI, Enterprise Applications, Data, Core Engineering, and Finance. Chennai / GDC Chennai 241596: .NET + Azure + Angular | 10–15 Years | Up to ₹25 LPA 242729: Salesforce Technical Lead 243106: Salesforce Marketing Cloud Specialist 243140: ELK Stack Developer 243552: SAP S/4HANA EWM Consultant 243941: SAP ABAP CDS Developer 244146: Full-Stack Engineer (Java, Spring Boot, React, PostgreSQL, GCP) 244508: Workday Extend & Orchestration Developer 244509: Senior Workday Integration Consultant / Developer 244559: Senior Workday Core HR Developer 244594: AWS Connect + Claude AI Developer 244595: ETL QA Analyst 244592: Endpoint Systems Engineer (SCCM, JAMF, Intune) Hyderabad 242505: Product Owner 242573: Guidewire Technical Architect (Policy & Billing) (Hyderabad / Bengaluru) 243548: Oracle Fusion Technical Architect (OIC, BIP, VBCS) 243547: Oracle Cloud Order Management (OM) Functional Architect 244459: AI/ML IoT Quality Engineer (SQA) 244623: Senior ETL Developer 244627: Business Analyst (Life Insurance Domain) Bengaluru / Bangalore 243833: ServiceNow Architect 244063: SME – SAP Tax 244066: SME – SAP VIM 244069: SME – SAP Vistex 244071: SME – SAP BTP Administrator 244527: C1 Analyst – IT Cloud COE 244676: IBM Planning Analytics Developer Mumbai 244271: AUM – Associate Director Level 244274: Junior General Corporate (GC) Accounting 244277: Fixed Assets & Billing Accounting Associate 244330: Oracle Fusion Cloud EPM/EPCM Functional Consultant (Cost Allocation) Pune 244276: Mid-Level Corporate Accountant 244275: Senior International Accountant Abu Dhabi 242512: AI Engineer 244424: SAP VIM Consultant 244693 / 244694: Forward Deployed Engineer 244695: AI Value Architect Dubai 244619: Full-Stack React Developer with .NET Backend (Airlines Domain | 5+ Years) 244621: Operations Specialist – Cargo Application Support Multi-Location Openings 244393: AWS Data Engineer (Chennai / Hyderabad) 244726: Databricks ETL Engineer (Bengaluru / Bangalore) How to Apply: Send your updated CV directly to chetan@kiashsolutions.com. Mention the Job ID and Designation in the subject line for immediate review. Know a rockstar candidate? Tag them below or repost to boost your network's career moves. #Hiring #TechJobs #CareerOpportunities #CloudComputing #DataEngineering #AIJobs #Salesforce #Workday #SAP #Oracle #FullStack #HyderabadJobs #BangaloreJobs #ChennaiJobs #MiddleEastJobs #RemoteJobs

Hello everyone,a run down from where I stopped. From Time Management to Data & Research Skills It’s been a while since I shared updates about my ongoing program, so here’s a quick run down: We kicked off with Time Management & Productivity with our amazing tutor, Mrs. Victory. Key takeaways: Key components of a remote work system Popular productivity methods, Time blocking, Batch working, Pomodoro technique The Eisenhower Matrix*: how to prioritize tasks + manage multiple clients Avoiding burnout while staying consistent Then we moved into Data & Research Skills and this really opened my eyes. I now see why it’s a core VA function. It drives client decision-making, boosts market value, and even helps with lead generation and income. This session was fully practical. We worked through real scenarios using different tools and AI tools like Perplexity AI, Browse AI, Claude, Google Gemini and more. Special appreciation to @phtechexpo and @IgniteVA for this opportunity and for making learning so hands-on Tomorrow is another day to go deeper into the VA Scholarship Program with IgniteVA. Ready for more growth! #VirtualAssistant #IgniteVA #PHTechExpo #Productivity #DataResearch #RemoteWork #LifelongLearning

Stop overcomplicating your marketing stack. You don't need 20 expensive tools to see real results...you just need a few that actually work together. ​When you use too many tools, you spend more time managing software than actually growing your business. ​Here is a simple, no-nonsense setup for modern marketing: ​AI for Speed: Use tools like ChatGPT or Claude to brainstorm ideas and outline content fast. (Just add your own human voice before posting). ​Quick Visuals: Use tools like Canva to create clean, simple graphics in minutes without waiting days for a designer. ​Smart Automation: Use Zapier or Make to connect your forms directly to your email list so you never copy-paste lead data manually again. ​Basic Analytics: Focus on native dashboard insights to see what actually brings in customers, not just vanity likes. ​The bottom line: Tools don't build a strategy...they just speed up your execution. ​Keep your setup simple, save your time, and focus on delivering value to your audience. What is the one marketing tool you rely on every single day? Drop it below ​#Marketing #DigitalMarketing #MarketingTools #Productivity #SimpleMarketing

UST is Hiring | Sr. Lead #BackendDeveloper | 5-7 Yrs | Kochi, Trivandrum, Hyderabad Apply or Refer: https://lnkd.in/gtcJZUST Skills: #Nextdotjs, #Nodedotjs, #REST, #SQL 📌 Note: ⚠️ Please apply only if your profile matches 80–90% of the requirements. Required Skills & Experience • Bachelor's degree in Computer Science, Engineering, or a related field. • 8+ years of backend software development experience, including 2+ years leading a team or squad technically. • Deep expertise in TypeScript and Node.js in production. • Strong experience with tRPC or a comparable typed RPC layer — routers, procedures, middleware, and the typed client/server contract. • Strong experience with Next.js server-side — route handlers, App Router API design, instrumentation.ts, and standalone output. • Zod or equivalent runtime validation at the API boundary. • Deep PostgreSQL — schema design, indexing strategy, query plans, and transaction isolation. • You should read EXPLAIN ANALYZE comfortably. • Experience with multi-schema PostgreSQL — search paths, cross-schema foreign keys, and permissions. • Strong ORM experience — Prisma and/or Drizzle — including migration strategy and the judgement to run two ORMs safely side by side. • Node streams / Web Streams — genuine understanding of backpressure, not just piping. • Dependency injection in a JavaScript/TypeScript runtime — Inversify or equivalent — including lifecycle and scoping in a serverless runtime. • Layered architecture discipline: thin routes, orchestrating managers, domain engines, persistence-only repositories. • Authentication and authorisation at implementation level — OAuth2/OIDC, JWT, session management, and permission modelling. • Application security fundamentals — OWASP Top 10, SSRF defence, secrets management, dependency advisory remediation. • Testing with Vitest, including database-backed integration testing. • Practical, day-to-day use of AI-assisted engineering tools (GitHub Copilot, Claude, ChatGPT or similar). • Experience working in Agile/Scrum environments.

🚀 Struggling to maintain prompt performance after migrating to a new AI model? Or building a complex agentic workflow from scratch? I have recently revisited an excellent presentation from earlier this year by Margot van Laar at Anthropic, where some of the practical strategies for optimising LLM prompts were shared. Transitioning from "vibes-based" prompting to rigorous system engineering is essential, particularly when maintaining production prompts during migration or building from zero to one. Here are some of the battle-tested prompt optimisation techniques to build reliable, high-performing AI systems: 🎯 1. Lead with Rigorous Evaluations (Evals) Never guess your improvements. Before you edit a single word of your prompt, build an eval suite to prove changes correlate to performance and prevent regressions. Cover control cases, edge cases, and refusal/handoff boundaries. 🧹 2. Practice Clean Prompt "Hygiene" Separate instructions cleanly into XML tags like <role>, <policy>, and <guidelines>. Ditch copy-paste fluff, false roles, and website remnants. Define a strict output contract using structured outputs paired with API stop sequences. 🛡️ 3. Stop "Patching" and Start Version Controlling Prompts often carry legacy patches written for older models. Because newer models excel at instruction following, they can overfit to these old guardrails and withhold valid information. Track prompt history to confidently retract outdated defensive fixes. 🛠️ 4. Give Models Capabilities, Not Just Loud Instructions Telling a model "it's critical to calculate correctly" does not make it better at mental math. If a task requires systematic computation, provide a tool schema. When defining negative constraints, present both sides of the trade-off so advanced models can balance them intelligently. 🤖 5. Go Agentic with "Generate-Evaluate-Repair" Loops Break down monolithic prompts. Split tasks into simple, independent prompts running in a loop: one to generate a draft, a second to evaluate for rule violations, and a third to execute targeted fixes. This is faster, uses fewer tokens, and supports soft runtime constraints without backend code modifications. Prompt engineering has transitioned from an art form to a systematic engineering discipline. If any engineering group is navigating this shift, this methodology is a highly recommended watch. Curious to know what evaluation framework are you currently using for your agentic workflows? 👇 Link to the full presentation in the comments. Amit Prakash Singh Vivek Singh Ruthambharaa T Hebbar Aditya Pratap Srinivas Ganumalla Arun V Sausthav Bora Manish Nainwal Neeraj Patankar Shivanand Gujjar Debashish Panda Tarni Sharma #PromptEngineering #GenerativeAI #Anthropic #Claude #AIAgents #EnterpriseAI #AIArchitecture

Most businesses are still losing thousands of dollars every month to missed calls and after-hours leads. 📞🤖 Yesterday, I hosted a live AI Founder Hub masterclass breaking down the end-to-end blueprint of how to build and sell production-grade AI Voice Call Assistants for $2,000+. This wasn’t just theory—we built a real-world, working solution live on the call. Here’s a quick recap of the roadmap we covered: 1️⃣ Market Opportunity & Valuation: How to position AI receptionists by highlighting the exact revenue clients lose from missed calls, staff overhead, and limited operating hours. 2️⃣ Automated Lead Generation: Scraping high-ticket decision-makers (CEOs/Founders) in seconds using Claude & Clay integrations. 3️⃣ Building the AI Voice Agent: Developing a multilingual, low-latency receptionist on Retell AI with custom system prompts, LLMs, and natural background audio. 4️⃣ Calendar & Workflow Automation: Integrating Cal.com and Make.com so the AI can check live availability, book meetings on the spot, and instantly trigger confirmation workflows. 5️⃣ The Client Acquisition Demo: How to set up an interactive instant-call demo form that converts prospects on the first try. Whether you are: 🔹 Building your own AI systems & agency 🔹 Looking to learn hands-on, practical AI automation 🔹 Passionate about teaching & sharing your AI expertise with a growing community 👉 Watch Full Master Class Here On My YouTube Channel: https://lnkd.in/dTCKJASV 👉 Join us and get access to all the blueprints, templates, and resources: 🔗 aifounderhub.com Let’s build together. 🚀 #AIAgency #VoiceAI #RetellAI #AIAutomation #ClaudeAI #MakeAutomation #AIFounderHub #LeadGeneration #ArtificialIntelligence

Claude vs ChatGPT: Is it really a one-sided game? 100 crore users vs 3 crore users. Sounds one-sided, right? I compared Claude and ChatGPT across 12 rounds, and the result wasn't as simple as the numbers suggest. Round 1: Basic Features Draw. Web search, PDF analysis, memory, Gmail & Drive integration, coding, and document creation — both handle these well. Round 2: Image Generation Winner: ChatGPT. Posters, social media creatives, marketing visuals — ChatGPT currently has the edge. Round 3: Voice Mode Winner: ChatGPT. For natural voice conversations and voice-based workflows, it's still ahead and more affordable. Round 4: Free Plan Winner: ChatGPT. Claude's free plan is more limited, especially for students and casual users. Round 5: Usage Limits Winner: ChatGPT. Many Claude users hit limits quickly, even on paid plans. Round 6: Affordable Entry Winner: ChatGPT. The Go plan offers a lower-cost starting point. Round 7: Learning Resources Winner: ChatGPT. With over 100 crore weekly users, tutorials, guides, and community support are everywhere — including in Bangla. Now let's look at the other side. Round 8: Writing Style Winner: Claude. Less flattery, more direct feedback. It often feels more like an editor than a cheerleader. Round 9: Coding Winner: Claude. Claude Code provides a strong developer-focused workflow. Round 10: Specialized Workflows Winner: Claude. Skills, persistent work instructions, Cowork, and Claude Design create unique productivity experiences. To be fair, Claude wins these rounds partly because ChatGPT doesn't offer identical features. That's less about "winning" and more about being different. Round 11: Ads Winner: Claude. Anthropic has publicly stated that Claude will remain ad-free. ChatGPT now shows sponsored content to some Free and Go users in supported regions. Round 12: Bangla Language No verdict. There isn't a reliable benchmark yet. Most opinions are based on personal experience. So this round belongs to you. My practical recommendation: Images, voice, quick questions, brainstorming → ChatGPT Long-form writing, coding, document analysis, repetitive workflows → Claude Making Reels? Generate ideas and thumbnails in ChatGPT, then write the script in Claude. Start with the free versions of both. Upgrade only the one that actually becomes part of your daily workflow. Two billion-dollar companies are competing aggressively, and users are getting access to incredible tools at a fraction of their value. ChatGPT leads on score. I still keep both open. Now it's your turn: Which one writes better in Bangla — ChatGPT or Claude? And which round's verdict do you disagree with? Save & Share. #Mergeasy #AIinBangla #ChatGPT #Claude #ArtificialIntelligence #Productivity #AITools #BanglaTech

Tu sitio puede cambiar mientras tu medición queda atrapada en la versión anterior ⚡ Una migración o un rediseño puede modificar URLs, formularios, botones y rutas de navegación. El sitio continúa funcionando. Pero Google Tag Manager puede seguir buscando elementos que dejaron de existir hace meses. El problema es que estas fallas rara vez generan una alerta visible. Con la IA y las APIs, el trabajo cambia. 𝗖𝗹𝗮𝘂𝗱𝗲 𝗖𝗼𝗱𝗲 / 𝗖𝗢𝗗𝗘𝗫 pueden conectar el sitio actual, GTM, GA4 y Google Ads para: ↳ Comparar lo que mide el contenedor con lo que realmente existe en el sitio ↳ Detectar URLs obsoletas, triggers frágiles y etiquetas huérfanas ↳ Reconstruir funnels y eventos sobre la navegación actual ↳ Medir formularios únicamente después de un envío exitoso ↳ Preparar y verificar las nuevas conversiones antes de publicarlas El agente inspecciona, cruza información y construye el plan de implementación. Las decisiones sobre qué medir y cuándo publicar siguen siendo humanas. No se trata únicamente de auditar un contenedor más rápido. Se trata de evitar que las campañas optimicen durante semanas con señales incompletas sin que nadie lo advierta. Eso es 𝗪𝗼𝗿𝗸𝘀𝗽𝗮𝗰𝗲 𝗩𝟱: ✅ Diagnóstico cruzado entre plataformas ✅ Medición reconstruida sobre el sitio real ✅ Señales confiables para optimizar Menos conversiones invisibles. Más datos reales para tomar decisiones. 🆕 Ya disponible en 𝗔𝗱𝗗𝗮𝘁𝗮 𝗔𝗰𝗮𝗱𝗲𝗺𝘆 · -·-·-·-·-·-·-·-·-·-·- · Tu futuro se construye con conocimiento 👉 Súmate a la evolución en AdData.Academy 🎓 #ClaudeCode #CODEX #PaidMedia #GoogleTagManager #GoogleAds #GA4 #MedicionDigital #AdDataAcademy #V5

Stellar finale to a stellar rox.com/teams launch week. 8 of the craziest things RevOps + GTM leaders are automating… from our Revenue Builders event! 1) Steve built an entire RevOps command center in Rox: one personalized operating system he can continuously edit as the business changes. - Monitors high intent signals → sends prospects personalized outbound - Delivers a targeted view of sales pipeline, out-quarter health, and areas of focus to drive across GTM - Surfaces biggest deal risks, recommended next actions, and everything else he needs to run RevOps. 2) Romain from OpenAI showed how powerful Codex computer use has gotten. We watched it complete an entire questionnaire in real time. It filled every field, clicked through every save + submit, and finishing the workflow end-to-end in a matter of seconds. 3) Bruno from XBOW built an entire GTM engine that includes... (and more) - An expansion-readiness agent that scores every customer on likelihood to expand, then builds a thesis for how that expansion should actually happen. - A community-listening agent that continuously finds people talking about XBOW across Reddit, social, and the web - A customer triage system that can take hundreds of incoming pings and surface the 5 that matter most directly in Slack 4) Jordan from Clay showed what happens when enrichment + personalization become one automated flow. A lead comes in, gets researched and enriched, and can immediately receive personalized memes, documents, and outreach without a human touching the process. The biggest theme: agents are moving way beyond helping with individual tasks + starting to run entire parts of the revenue cycle. That’s exactly what we built Rox for! Rox.com

Yesterday I asked a Head of RevOps at a 20-person AI startup what GTM stack they're running. His answer: no GTM SaaS at all. Custom backend, Postgres for storage, a durable workflow layer for orchestration, agents doing the research. Data providers are just API endpoints they swap in and out. A year ago most RevOps people would have had no idea what these things even mean, and probably most still don't. Of course there are some downsides to building out a setup like this: technical maintanence, debugging, errors, etc. But there's also a huge upside from flexibility, cost, and ability to get creative with GTM plays. Based on how well companies like Salesforce, HubSpot, and Attio are building out their headless functionality, I wouldn't recommend this for more established companies but you should 100% be using Claude/Codex to orchestrate across your tech stack and drop all the tools that don't have APIs available. They're just going to slow you down. I think we'll start seeing more AI native companies building out their tech from scratch like this. It's too fun what you can do when you can control you're entire GTM from a terminal :)

OpenAI’s expansion in Brazil made me think about a question: How should a global AI product enter a local market? The traditional playbook often looks like: Choose a market → Localize → Launch campaigns → Acquire users But Brazil suggests another path: Localize enough → Observe organic adoption → Follow product pull → Invest deeply 1️⃣ Minimum localization makes adoption possible Organic adoption does not mean localization is unnecessary. Users still need the product to be accessible and usable in their own language. But there is an important distinction: Making a global product locally usable ≠ deeply investing in a market Language localization opens the door. It doesn’t tell you how deeply to enter the market. Once the basic friction is removed, real user behavior can start revealing where demand already exists. 2️⃣ Organic adoption reveals where the product pull is Brazil was already showing strong adoption before OpenAI established its local team in São Paulo. According to OpenAI, Brazil is now: ▸ A top-three market globally for ChatGPT weekly active users ▸ Generating around 215 million messages per day ▸ The second-largest market for OpenAI API developers ▸ The largest Codex market in Latin America At this stage, the GTM question is no longer: Will Brazil want ChatGPT? It becomes: How do we deepen a market that already wants it? Organic adoption becomes more than a growth metric. It becomes a form of market research. 3️⃣ Local behavior tells you how to localize The next step is not simply translating more content or spending more on local campaigns. It is understanding how local users are actually using the product. Brazil, for example, is one of the strongest markets for ChatGPT Images and also shows high Voice usage. OpenAI’s local expansion also includes initiatives such as a São Paulo Creators Day, Portuguese-language developer programs, hackathons, education initiatives, enterprise engagement, and public-sector partnerships. This is what I find most interesting: Localization can become behavior-led. Instead of applying the same GTM playbook to every country, companies can use local usage patterns to decide which users, use cases, communities, and ecosystems deserve deeper investment. ✨ For global AI products, organic adoption can become a signal for where to invest — and local behavior a guide for how to invest. Sometimes the users choose the market first — and their behavior tells you how to enter it. #OpenAI #GlobalGTM #AIProduct #ProductMarketing

Claude code’s performance has fallen off a cliff recently, and everyone seems to be talking about it on here. But I think I know what happened… Claude Code is one of the main tools in my tech stack (probably second only to Clay) and something I use everyday. Now I haven’t been using it for software dev but mainly for gtm stuff like editing Smartlead campaigns through the API, manipulating / joining datasets, scraping the web etc. But then a couple weeks back it started going wild. I would give it a fairly basic task and it would go off on random tangents running side quests that were completely unnecessary and give long complicated outputs. Put simply: burning credits and wasting time. I was on the cusp of switching to codex until I brought it up in the StackOptimise ⚙️ weekly GTME call, and the AI wizards Done Miladinov and Muhammad Rafay came to the rescue. “Yeah Opus 5 is sh*t. Switch back to 4.8” I hadn’t even realized the model had changed. Lesson learnt. So I did and the Claude I know and love returned. Anyone else noticed this lately?

We´re still looking for Senior GTMEs based in Pakistan, you can make $5k-$6k if you have: - At least 3 years agency experience - Built, managed and scaled intent driven Outbound. - You MUST very good at Clay, extra points if you are using Claude Code or Codex to run your daily GTM ops. If you don´t have any of the above, please don´t apply. There are no openings for juniors atm. Application link is in the first comment.

At Rence (YC F26), everyone from GTM to engineering contributes to the codebase weekly. A year ago this would be reckless. Today it’s just the optimal play, and has been the most insane productivity unlock we’ve had so far. Speed is everything at our stage and tokens are abundant, so leaving them unused is a strictly worse decision than letting someone try something. Pre-AI, engineering-orgs happily let junior engineers ship into production, not because they were always right, but because they learn fast and someone reviews the output. A non-engineer + Fable 5 is arguably 8.75x this today (trust), and has better conditions to learn, fast. None of this works without engineers, of course. More people shipping means more review, and the fix isn’t to review harder, it's building an environment where agents are more likely to get it right the first time. Anyways, if anyone has an idea, they should be allowed to execute, and below is an image of our beloved Codex executing our team.

One viral AI prompt burned $1,700 in tokens. It's called the gauntlet loop, and it's the best prompting idea of the year. Builder agents make the work. A blind critic compares it against a real reference. Anything that loses goes back for another round. Matt Shumer used it to build a Call of Duty clone from three paragraphs of text. 55,000 lines of code, zero hand-made assets. The internet lost its mind. Then the bills arrived. $1,200 for an F1 game. $1,700 for a GTA attempt: 22 hours, 86 agents. The top Reddit comment on the whole trend: "guaranteed token burn with fingers-crossed results." The technique is brilliant. The spending model is unsustainable for most. So I wanted to put my own spin on it with spending guardrails. Same loop, one new rule: it refuses to start without a budget. You give it 50% of your usage window and it sizes the entire run to fit. Runs out? It stops, shows you what the spend bought, and asks before touching more. 50, then 75, then 95. Never on its own. I tested it with the same one-line brief, same model, run twice. The naive single pass shipped a landing page with a fake "as seen in Wired and TechCrunch" press bar. Completely invented. Nothing in a single pass ever asks "says who?" The gauntlet build couldn't get away with that. Its critic rejects any claim that can't point to evidence. Contrast ratios computed. Keyboard actually pressed e2e. Every number traceable to the brief. GTM teams, this is where it gets useful for you. Point it at a landing page, a competitor battlecard, a launch email sequence, or a case study. The critic checks marketing claims the same way it checks code: no invented stats, no fake logos, no "as seen in" that legal never approved. brief becomes the quality bar. What ships is what passed. Free and MIT licensed. Works on Claude Code, Codex, Hermes, or agent of your choice. Install is one command. Link in the first comment. Try it out, if you like it give it a star! Real question for anyone running agent fleets: do you cap your runs, or let them cook?

Our SaaS had a massive problem in august. Only 6.25% of our signups got value out of our product. Welcome to "Build OXYGEN in public" week 1, where I share our weekly learnings, f*ckups and wins. So here is where we screwed up: [1] We asked for a credit card on the trial period 50% dropped off right there. So we essentially invested time, money and energy to acquire a user and 50% of them just droped off on a pay wall. [2] Activation was 12.5% Of the half that got past the credit card wall, 1 in 8 reached value (which we measure in wether the user has set up a workflow, sequence or table in our product). 50% x 12.5% = 6.25%. 1 in 16 signups. [3] Churn was 13% 13% monthly churn means we replace our entire customer base every 8 months. It's well known that GTM tech has high churn, but this is way too high. [4] Support was a disaster We launched and people texted us on LinkedIn, Slack and WhatsApp. Tickets ran through a vibecoded system in our CLI and MCP 🫠 So here what we shipped to solve all of that: [1] A free tier, no credit card Plus 10$ in credits gifted on every signup. So you can find out if it works before you pay us anything. [2] A simpler onboarding Less steps between signup and your first table, sequence, workflow. We also invested time into creating some templates which can be used as a base. [3] An in app AI copilot It builds the setup for you if you don't want to open a terminal. (I still believe everybody should run this through Claude Code or Codex, but that can't be the only way in). [4] Plain as our support system Slack, in app chat and an API based option in one inbox. Makes it much easier for us and our agents to help out our customers. So here are my goals for the next 30 days: - activation from 12.5% to 25% - churn from 12.5% to 8% I ranted a lot in this post, but there has been massive wins too! We stopped our entire marketing and sales engine to fix our leakages first, but we still grew our MRR and signups. The customers that use us, love the product and tell others about it. Now that we make our product much more intuitive we are ready to go all in on GTM again heheh. So follow Tim Scheuer to read about our progress next Friday.

GTM Engineers - your in real danger of falling behind if you don't evolve quickly: A few weeks ago, Clay released Workflows. In one swoop, they changed GTM orchestration forever. Think back a few years ago when Clay released tables. They took CSV files, and brought them to life in the Cloud. Microsoft, in all its billions, didn't figure out how to activate a table and a row, a concept they invented, yet Clay did. Now here we are in 2026. More people use Claude, OpenAI, Cursor, etc. All Clay's customers. So what did Clay do, build a managed layer for agents. They entered Phase 2 of GTM dominance. In doing so, they found the gap that was missing in GTM. No limit executions, python code (think cleaning, classification, etc), opened up the API to build on top of it for customers. I'm not joking when I say they swallowed the 5 year strategy of Zapier, N8N, and Make in one swoop. How do I know, because I build stuff quick. Something new comes along, I try it. There is nothing I can't build now using workflows. I'll never build a table again. Save this post, because in a year, you will look back and think, glad I listened. Governed GTM at scale, from the AI of you choice has arrived, the only question will be, did you miss it. PS: pardon my Codex pet, he refused to leave the image:)

Most companies use AI to create more GTM activity I think the bigger opportunity is to make every campaign teach you something that improves the next one Imagine you’ve already run 30 campaigns For each campaign, you know who you targeted, which problem you thought they had, what information you used, what you sent, and whether it produced meetings or revenue The problem is that this information is usually spread across different tools, or simply forgotten when the campaign ends What if you kept a simple record of every campaign? Then AI could compare all of them and help you answer questions like: Which problems actually led to revenue? Which information helped us choose the right companies? Where did the results match the ROI we promised? Which campaigns looked good but produced nothing? The ROI part matters because a prospect should understand what they could get from working with you and why you believe that result is possible Suppose you sell software that helps finance teams finish their monthly reporting faster “AI for finance teams” doesn’t say much A better message could explain how much time similar companies saved, what slowed them down before, and which details you used to estimate the result for this specific company You’re probably asking, how do you build this? Start with 10 campaigns that worked and 10 that didn’t Write down the audience, the problem, the evidence you used, the result you promised, and what actually happened Give that information to Claude/Codex and ask it to find the patterns, then review those patterns yourself and decide which ones are worth testing again Your next campaign now begins with everything you learned from the previous ones Run the campaign, keep what you learned, and use it the next time After enough campaigns, you have your own record of what actually produces revenue, and that becomes very difficult for another company to copy

We’re hiring across the teams helping startups build, scale, and tell their stories with OpenAI. These are truly once-in-a-career opportunities working with the sharpest minds in AI (only a little biased)! Sharing a few roles that will work on/closely with our team & have an outsized impact in how we show up for builders around the world: Startup Content Marketing Manager (SF): Get close to founders building with OpenAI, uncover the stories and lessons behind their products, then turn them into high signal content that builders actually want to read, watch, and share. 👉🏼 https://lnkd.in/gVVcb_nN Integrated Marketing Manager, Developers & Startups (SF): Bring large-scale brand campaigns to life, shaping the strategy, briefing agencies, landing the creative, and making sure builders are at the center of the story. 👉🏼 https://lnkd.in/gQJnX-Hw Growth Programs Lead, Startups (SF): Build the operating system that helps startups grow with OpenAI by turning credits, partnerships, and GTM motions into scalable, measurable programs. 👉🏼 https://lnkd.in/gJbkM2JA Account Director, Startups (Global): Be a trusted partner to founders. Understand what they’re building, anticipate what they need next, and help them get more value from OpenAI at every stage of their journey. (Fun fact, 80% of our Startups GTM team are former founders!) 👉🏼 https://lnkd.in/gGmmCjPk Resource Manager, Business Marketing (SF or Remote): Help our team move faster and smarter by matching the right work to the right people, simplifying how projects get prioritized, and building AI-powered workflows (we love Codex). 👉🏼 https://lnkd.in/g23-J5Cj // Please note I won't be able to respond to DMs but recommend applying directly as roles are filled quickly. All open roles can be found at openai.com/careers.

The best way to learn GTM Engineering? Build something you actually need. I've realized that reading about AI agents, automation, and GTM workflows is useful. But nothing compares to trying to build one. You quickly discover: → Your data isn't as clean as you thought. → Your workflow has more edge cases than expected. → Your prompt isn't as reliable as you expected. → The automation breaks somewhere you didn't anticipate. And that's exactly where the learning happens. That's why over the next few months, I'm trying to spend less time asking: "What can this tool do?" And more time asking: "What can I build with it?" Claude. Codex. n8n. Vibe-coding platforms. AI agents. I'll be experimenting with all of them to build practical GTM systems and small products. Some will work. Some definitely won't. 😄 I'll share both. Build → break → learn → rebuild. That's the journey. 💬 For everyone building with AI: What's the most useful thing you've built for yourself recently? #GTMEngineering #AI #GTM #AIAgents #Automation #VibeCoding #RevOps

Team-led content is cheat code for B2B growth in 2026. Companies are generating millions with exactly this strategy ↓ (Save it before it gets lost in the feed.) Most companies still create content like this: Open a blank AI chat. Write a prompt. Fix the generic draft. Repeat everything tomorrow. Here is how best teams operate. They create a folder which claude or codex can access with: 𝟭. 𝗕𝘂𝘀𝗶𝗻𝗲𝘀𝘀 𝗯𝗿𝗮𝗶𝗻 ICP, positioning, messaging, offers, proof, objections, customer language, and the founder's point of view. 𝟮. 𝗥𝗲𝗳𝗲𝗿𝗲𝗻𝗰𝗲 𝗹𝗶𝗯𝗿𝗮𝗿𝘆 The best posts, hooks, infographics, carousels, and images are saved as quality benchmarks. The system studies the patterns without copying the work. 𝟯. 𝗣𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻 𝗽𝗹𝗮𝘆𝗯𝗼𝗼𝗸𝘀 Step-by-step instructions for posts, carousels, infographics, reviews, and repurposing. HTML templates keep every visual consistent and editable. 𝟰. 𝗜𝗱𝗲𝗮 𝗲𝗻𝗴𝗶𝗻𝗲 Reddit threads reveal pain. X shows emerging conversations. Competitors reveal crowded angles. Sales calls provide buyer language, objections, and proof. 𝟱. 𝗥𝗲𝘃𝗶𝗲𝘄 𝗹𝗼𝗼𝗽 Every draft is checked against the ICP, positioning, voice, proof, hook quality, and pipeline outcome before it ships. Now a team member does not need to guess what the founder means. They open the workspace, follow the playbook, and improve the shared system with every post. AI handles execution. The team's accumulated context protects the quality. What is your content creation workflow with AI? ♻️ Repost to help your founder friend nail GTM ➕ Follow Kanchan Bhatta for GTM systems My Tool Stack: Claude Code Cursor Google Workspace Reddit, Inc. X Higgsfield AI #GTM #B2BSaaS #FounderLedGrowth

GTM Engineers, you have to understand how everything works, its your job! By now, we all heard of ChatGPT, Gemini, Claude. But have you used Codex, Claude Code, or Cursor. Did you know Cursor is owned by SpaceX and because of that OpenAI is leaving in November. Do you understand Open Source, have you heard of GLM, are you familiar with open weights. Can you install a CLI or MCP? Its a lot, I know. But as a GTM, you are expected now to be just as much an AI Engineer at you are a GTM Engineer. Are you testing Grok and Devin. Did you know Grok has its own computer, yet most work you do in Codex is on your local machine. Your boss will expect you to know how data will be protected, what is being shared and most importantly, what does everything cost. You need to be able to explain why you need what you need to do your job. Your business has a budget, and unless you work for a unicorn, its not unlimited!

GTM skills are evolving fast. Yesterday, I got to build one. At the GTM Skillathon in Bucharest, together with Lidia Mititelu , we built a skill focused on AI Search Optimization and online visibility. For our case study, we chose SalesOMMO. The goal was simple: understand where they currently stand online, who they compete with, what search opportunities exist and what they could improve. Because we only had around 2 minutes of execution time, we had to be very intentional about the workflow. Instead of making the agent search for everything from scratch every time, we first used Apify to extract the relevant web and search data, then fed that context directly into the agent before running the analysis. That made the workflow faster and gave the agent better context from the start. From there, the skill could compare the business with its competitors and turn the data into clear recommendations for improving online visibility. That was probably my favorite part of the challenge, not just building something with AI, but figuring out how to make it actually useful under real constraints. And honestly, the community was just as important as the project. A room full of people building, testing ideas, helping each other and shipping in a few hours is exactly why communities like Builders House are so valuable. Huge thanks to Alexandru C. ,Formidable Builders, Builders House, OpenAI and Apify for making it happen. And congratulations to all the winners, amazing gtm skills !🥇 And, of course, thanks to Lidia Mititelu for building this with me. Github repo: https://lnkd.in/d5DSYTVH Fun fact: we pushed our last commit 30 seconds before stop coding… so naturally, we only managed to take one photo together 😂 #BuildersHouse #FormidableBuilders #AgentSkills #AI #Codex #OpenAI #Apify #GTM #SEO #BuildInPublic

We had the best time yesterday at GTM Skillathon. ⚡️ Together with the one & only Radu Minea, I enjoyed every moment building a skill for OpenAI Codex. And the best part was that we met so many talented & insightful people, you can check out all the projects in the public repo. Thank you, Formidable Builders, Builders House & everyone involved. See you soon! 👀

Your GTM workflows should be in github right now. maybe not a hot take, but i just got off a call where everything was either in spreadsheets or siloed in 6 different tools. makes no sense. grab your API keys. build a knowledge base with claude, codex, whatever agent you use. version control it. then you just ask: what’s my deal state? what were the reply rates on the last campaign? no digging through 500 tabs. one answer, with evidence, that tells you how to make the next campaign better. Drake gets it 🤣

Startup Idea or Hackathon Idea " DriftForge " Codex - Antigravity - Claude Code - Grok Build - Cursor - Replit prompt " Build a web app called DriftForge that helps GTM teams monitor long-running sales agents for eval drift and production regressions. Features: an upload area for mocked support tickets, sales traces, and human labels; an eval suite dashboard showing pass rates by task type; a drift detector comparing recent failures against older labeled cases; a prompt-change timeline with impact notes; and a review queue where humans approve new golden examples. UI: modern operations dashboard with trace cards, trend lines, and high-signal alerts. Start with mocked data."

Now that OpenAI - Codex is watching what I'm doing with computer history, here is what it says I should post. I mean, I did use computer history and had it review my analytics. We will see how accurate it is: Today I told an AI agent to stop finding leads. It had looked at a 5,400-person campaign, found 2,115 verified records inside one review set, and confidently concluded we needed 3,285 more. The math was clean. The conclusion was wrong. Another 2,676 people were already sitting in the campaign audience. They had not been reconciled yet. If I let the agent continue, it would have sourced duplicates, spent money we did not need to spend, and made the campaign harder to trust. This is the part of agentic GTM the demos skip. Finding 5,000 people is easy. Knowing the difference between existing, pending, verified, review-held, uploaded, and approved to send is the actual work. So I stopped sourcing and changed the order: 1)Reconcile what exists. 2)Verify what is unresolved. 3)Read back the provider state. 4)Then decide whether a real gap remains. AI does not become operational because it can take action. It becomes operational when it knows which actions it is not allowed to take. That is the difference between an automation and an operator.

Stellar finale to a stellar rox.com/teams launch week. 8 of the craziest things RevOps + GTM leaders are automating… from our Revenue Builders event! 1) Steve built an entire RevOps command center in Rox: one personalized operating system he can continuously edit as the business changes. - Monitors high intent signals → sends prospects personalized outbound - Delivers a targeted view of sales pipeline, out-quarter health, and areas of focus to drive across GTM - Surfaces biggest deal risks, recommended next actions, and everything else he needs to run RevOps. 2) Romain from OpenAI showed how powerful Codex computer use has gotten. We watched it complete an entire questionnaire in real time. It filled every field, clicked through every save + submit, and finishing the workflow end-to-end in a matter of seconds. 3) Bruno from XBOW built an entire GTM engine that includes... (and more) - An expansion-readiness agent that scores every customer on likelihood to expand, then builds a thesis for how that expansion should actually happen. - A community-listening agent that continuously finds people talking about XBOW across Reddit, social, and the web - A customer triage system that can take hundreds of incoming pings and surface the 5 that matter most directly in Slack 4) Jordan from Clay showed what happens when enrichment + personalization become one automated flow. A lead comes in, gets researched and enriched, and can immediately receive personalized memes, documents, and outreach without a human touching the process. The biggest theme: agents are moving way beyond helping with individual tasks + starting to run entire parts of the revenue cycle. That’s exactly what we built Rox for! Rox.com

This is where AI becomes much more powerful. 🚀 It’s not about using one AI tool. It’s about connecting the right tools to create an AI-powered workflow. → Claude Code — terminal & development workflows → Codex — coding & automation tasks → Clay — lead generation & enrichment → Apify — web scraping & data collection → MCP — connecting AI to tools & data → n8n — workflow automation → LangGraph — building AI agents → Supabase — backend infrastructure The biggest shift is not learning more AI tools. It’s learning how to connect them. That’s when AI stops being a collection of individual apps and starts becoming infrastructure for the way you work. The future isn’t about asking, “Which AI tool should I use?” It’s about asking: “How can I connect these tools to automate the entire workflow?” That mindset can transform productivity, operations, sales, development, and decision-making. AI is becoming a connected system—not just a collection of tools. What AI tools are you currently connecting in your workflow? 👇 #AI #GenerativeAI #Automation #AIAgents #Productivity #n8n #MCP #AItools

Claude code’s performance has fallen off a cliff recently, and everyone seems to be talking about it on here. But I think I know what happened… Claude Code is one of the main tools in my tech stack (probably second only to Clay) and something I use everyday. Now I haven’t been using it for software dev but mainly for gtm stuff like editing Smartlead campaigns through the API, manipulating / joining datasets, scraping the web etc. But then a couple weeks back it started going wild. I would give it a fairly basic task and it would go off on random tangents running side quests that were completely unnecessary and give long complicated outputs. Put simply: burning credits and wasting time. I was on the cusp of switching to codex until I brought it up in the StackOptimise ⚙️ weekly GTME call, and the AI wizards Done Miladinov and Muhammad Rafay came to the rescue. “Yeah Opus 5 is sh*t. Switch back to 4.8” I hadn’t even realized the model had changed. Lesson learnt. So I did and the Claude I know and love returned. Anyone else noticed this lately?

We´re still looking for Senior GTMEs based in Pakistan, you can make $5k-$6k if you have: - At least 3 years agency experience - Built, managed and scaled intent driven Outbound. - You MUST very good at Clay, extra points if you are using Claude Code or Codex to run your daily GTM ops. If you don´t have any of the above, please don´t apply. There are no openings for juniors atm. Application link is in the first comment.

$5B valuation. 14,000 customers. $100M+ ARR. And they just handed their most technical users the exit. Clay shipped an API. Every serious observer said they never would, and the logic was sound: an API doesn't kill revenue directly, it kills stickiness — and lost stickiness kills revenue on a delay. Three things were holding that moat: → Aggregated access to 150+ data providers → Your templates, prompts and logic living inside their environment → An agency channel selling on their behalf An API threatens two of the three. So why do it? Because the migration had already happened. The most technical users — the ones spending the most — had stopped opening the UI. They were running the same workflows from Claude Code, Cursor, Codex. The API didn't open a door. It acknowledged that people were already climbing out the window. That's the part worth sitting with. A company at that scale looked at its highest-value segment and concluded the interface was no longer where the relationship lived. Better to be the thing the agent calls than the tab nobody opens. I think that's the right read, and I think it's about to happen across the category. Within a year, the vendors that treat the agent as a first-class user — real API, MCP server, bulk operations, machine-readable errors — take the most technical and highest-spending slice. The ones that don't become the tools a human has to babysit. Babysitting is what churn looks like six months before it shows up in the numbers. Which of your vendors could an agent actually drive today?

GTM Engineers - your in real danger of falling behind if you don't evolve quickly: A few weeks ago, Clay released Workflows. In one swoop, they changed GTM orchestration forever. Think back a few years ago when Clay released tables. They took CSV files, and brought them to life in the Cloud. Microsoft, in all its billions, didn't figure out how to activate a table and a row, a concept they invented, yet Clay did. Now here we are in 2026. More people use Claude, OpenAI, Cursor, etc. All Clay's customers. So what did Clay do, build a managed layer for agents. They entered Phase 2 of GTM dominance. In doing so, they found the gap that was missing in GTM. No limit executions, python code (think cleaning, classification, etc), opened up the API to build on top of it for customers. I'm not joking when I say they swallowed the 5 year strategy of Zapier, N8N, and Make in one swoop. How do I know, because I build stuff quick. Something new comes along, I try it. There is nothing I can't build now using workflows. I'll never build a table again. Save this post, because in a year, you will look back and think, glad I listened. Governed GTM at scale, from the AI of you choice has arrived, the only question will be, did you miss it. PS: pardon my Codex pet, he refused to leave the image:)

Good news! My mobile shop is open today at The Charlottesville City Market. Thanks to the Launch Pad Program and The Community Investment Collaborative that makes this possible for collaboration. Featuring my book “The Dakini Codex.” A book that explains complex form of science and spirituality and presents them in an easy to understand format with lyrical poetry. It has been featured at the world’s largest book signing event in Frankfurt, Germany. It has been nominated for The Eric Hoffer Award and placed in the museum at The University of Science and Philosophy in Waynesboro, Virginia. My crystal jewellery are Somatically attuned and reiki infused for wearable forms of healing art. Serenity Hope is featuring two new product lines! A crystal worry stone set in a marbelized polymer clay cabochon. It can be worn as a pendant and used as a fidget to ease anxiety. I also have upscale recycling pendants from bottle caps. They are a toast to your favourite drink! A beautiful example of how we can keep mementos of precious memories while being conscious of how we live in right relation to the planet and her resources. Come and visit! We love to see faces from the community. Together we make a difference. 🕉️🙏🏻

Most founders chase $1M ARR with Random tools. 𝗪𝗿𝗼𝗻𝗴 𝗺𝗼𝘃𝗲. The founders who actually hit that number don't use more tools. They use the RIGHT seven. (📌 𝗦𝗮𝘃𝗲 𝘁𝗵𝗶𝘀 before you build your next system.) ➡️Here's the exact stack that gets you there:👇 ☑️𝟭. 𝗟𝗢𝗩𝗔𝗕𝗟𝗘 Stop waiting on your dev team for every test. - Design a prototype in hours, not weeks - Launch and validate before you commit budget - Kill bad ideas early, cheap This is how you move at founder speed. ☑️𝟮. 𝗦𝗘𝗔𝗥𝗖𝗛𝗔𝗕𝗟𝗘 SEO used to mean guessing and waiting months. Now it runs on autopilot: - Tracks SEO, AEO, and GEO together - Adjusts before rankings even drop - Frees you from checking dashboards daily Your growth assistant that never clocks out. ☑️𝟯. 𝗖𝗨𝗥𝗦𝗢𝗥 Your engineering team is your biggest cost center. Cursor cuts that cost: - Writes code alongside your devs, not for them - Cuts review and debug time hard - Ships features faster without new hires Speed here compounds everywhere else. ☑️𝟰. 𝗭𝗔𝗣𝗜𝗘𝗥 Manual customer follow-up doesn't scale past 100 users. - Connects your CRM to every other tool - Automates retention sequences - Triggers upsell campaigns without a human touch This is the difference between managing customers and losing them. ☑️𝟱. 𝗖𝗛𝗔𝗧𝗚𝗣𝗧 + 𝗖𝗟𝗔𝗨𝗗𝗘 Generic onboarding loses new hires fast. - Train each one on your actual team voice - Build support docs that sound like you - Personalize training without hiring a trainer Two tools. One consistent brand voice at scale. ☑️𝟲. 𝗖𝗟𝗔𝗬 Sales teams waste hours enriching contact lists by hand. - Pulls social, sales, and marketing data into one view - Enriches leads automatically at scale - Hands your team warm context, not cold names Better data means fewer wasted calls. ☑️𝟳. 𝗖𝗟𝗔𝗨𝗗𝗘 𝗖𝗢𝗗𝗘 Churn kills ARR faster than any competitor does. - Flags churn signals before they show in your reports - Analyzes trends across your entire customer base - Automates the follow-up communication Catching churn early is cheaper than replacing revenue. Here's the pattern: ↳ Lovable → build fast ↳ Searchable → grow visibility ↳ Cursor → ship code ↳ Zapier → automate ops ↳ ChatGPT/Claude → train your team ↳ Clay → enrich your leads ↳ Claude Code → cut your churn Seven tools. Seven bottlenecks removed. Most founders think $1M ARR needs more effort. It needs fewer gaps in the system. 𝗥𝗲𝗽𝗼𝘀𝘁♻️ this if it helped ps: Which one of these are you not using yet? Comment below. ____ Nikhil

El trabajo de growth en 2026 no se parece en nada al de hace dos años. Antes todo corría sobre una lista corta de herramientas. SEO era Google. Paid era Meta y Google. Los datos salían de ZoomInfo. El sitio estaba en WordPress. Los reportes eran Google Analytics 4 y una planilla. Las automatizaciones, si había, eran Zapier. El outbound era mandar un mail desde Gmail. Ahora mirá sobre qué corre el mismo trabajo. Ya no te buscan solo en Google: también en ChatGPT, Claude y Gemini. Los reportes viven en PostHog y HockeyStack. El outbound corre en Instantly.ai, lemlist y HeyReach.io La automatización, en Clay, n8n y Make. El paid se expandió a LinkedIn, Reddit, Inc. y X. El sitio se construye con AI en Framer o Vercel. Pero el cambio real es otro. Es cuánto puede hacer una sola persona. Una idea del lunes a la mañana puede estar viva el mismo día. Investigar el mercado. Armar la lista de contactos. Publicar una landing. Hacer los creativos y lanzarlos. Y ver los primeros números a la tarde, sin esperar a nadie. Ese es el cambio grande, porque el trabajo viejo estaba lleno de dependencias. Entender un mercado nuevo eran días de research. Una landing nueva era esperar a un desarrollador. Datos limpios eran horas de trabajo manual. El stack nuevo se llevó casi todo eso. Y por eso cambia lo que te hace bueno en este trabajo. Ser excelente en un canal sigue contando. Pero hoy todos tienen las mismas herramientas, así que ahí la ventaja dura poco. Lo que separa es armar el sistema completo y ponerlo en el aire antes de que se pase el momento. Antes ganabas por saber más que el otro. Ahora ganás por llegar antes. --- 👉 P.S. Armar el sistema completo, y no solo dominar un canal, es lo que ordena el Curso Growth Rockstar: modelos de growth, adquisición, retención y monetización aplicados a tu negocio. Por ahí ya pasó gente de Rappi, Mercado Libre, Nubank y Tiendanube. La edición 15 arranca el 14 de septiembre y quedan los últimos cupos para aplicar. Comentá "GROWTH" y te mando el link

Probé decenas de herramientas de IA este año. Estas 10 son las que quedaron. Una por tarea, sin repetir función: 1 - ChatGPT. El todoterreno del día a día. 40 variantes de anuncio en 10 minutos, en vez de 4 en una tarde. 2 - Claude. Texto largo y análisis a fondo. 30 entrevistas a clientes convertidas en un mapa de objeciones con citas textuales. 3 - Gemini. Datos dentro de Google. Limpia, segmenta y puntúa tu base de leads sin sacarla de la hoja. 4 - Perplexity. Investigación con fuente al lado. Competencia, precios y quién decide, verificable en un clic. 5 - Clay. Prospección enriquecida. De 1.000 contactos genéricos a 80 con un motivo real para escribirles. 6 - Apollo.io. Base de contactos B2B. Correo, cargo y empresa de tu lista objetivo en minutos. 7 - n8n. Automatización de flujos. Conecta formulario, CRM y Slack sin escribir una línea de código. 8 - Gamma. Presentaciones y propuestas. Una propuesta de cliente lista en 15 minutos, no en dos días. 9 - Canva. Piezas gráficas. 20 creativos para tu próximo test de anuncios sin pasar por diseño. 10 - Fathom.ai. Llamadas y voz del cliente. Graba cada demo y te deja por escrito las objeciones que salieron. El error común es abrir las 10 al tiempo. Empieza por una: la que resuelva la tarea que más horas te come esta semana. Compartimos este tipo de contenido cada semana en La Liga del Growth, una comunidad de WhatsApp de operadores de growth en Latam: https://lnkd.in/eVeUUkNa

The AI winners of 2026 won’t use every app, they’ll build the right stack. Here’s the shortlist. 1 → General assistants → Claude pressure-tests strategy and long-form decisions. → ChatGPT analyzes files and ships routine work. → Perplexity researches markets with cited web answers. 2 → Development and building → Cursor edits entire codebases through prompts. → Replit prototypes and deploys products in one place. → Base44 builds internal tools without infrastructure overhead. 3 → Content creation → Beehiiv publishes and monetizes a founder newsletter. → HeyGen creates presenter videos without a crew. → Opus Clip repurposes long videos into social clips. → Manus AI handles multi-step research and execution. 4 → Productivity → Granola captures meetings without manual notes. → Gamma turns rough ideas into polished decks. → Superhuman triages email with AI-assisted workflows. → Grammarly tightens writing before customers see it. → Otio AI organizes and summarizes research projects. → Wispr Flow turns dictation into drafts faster than typing. 5 → Creativity → Suno generates original music for campaigns. → Canva creates branded visuals without extra headcount. → ElevenLabs produces natural voiceovers and localized audio. → Runway generates and edits campaign video. → Kling creates cinematic video from text or images. → Pika Labs animates ideas into shareable videos. → Figma moves designs into prototypes faster. → Google Veo produces polished scenes from prompts. → Higgsfield creates camera-controlled social video. 6 → Automation and integration → n8n connects tools through flexible workflows. → Clay enriches leads and personalizes outbound. → Claude Code delegates coding inside your terminal. → Softr turns operational data into client portals. → Gemini works across Google files and research. Which 5–6 tools would earn a permanent place in your AI stack? 𝗙𝗥𝗘𝗘 (𝗚𝗼𝗼𝗴𝗹𝗲) 𝗖𝗼𝘂𝗿𝘀𝗲𝘀 𝘆𝗼𝘂 𝘄𝗶𝗹𝗹 𝗿𝗲𝗴𝗿𝗲𝘁 𝗻𝗼𝘁 𝘁𝗮𝗸𝗶𝗻𝗴 𝗶𝗻 𝟮𝟬𝟮𝟲. 1. Google Data Analytics: https://lnkd.in/gmpYBGQz 2. Google Project Management: https://lnkd.in/g2SFqfdw 3. Foundations of Project Management: https://lnkd.in/gAphbnV7 4. Google Introduction to Generative AI: https://lnkd.in/gHnK6GAA 5. Google Cybersecurity: https://lnkd.in/gwT8n2XD 6. Google UX Design: https://lnkd.in/gtifc_sH 7. Google Digital Marketing & E-commerce:https://lnkd.in/gNUkiJN6 8. Google IT Support:https://lnkd.in/gHrnNNrc 9. Web Applications for Everybody:https://lnkd.in/gm5jjAEu 10. Get Started with Python: https://lnkd.in/gE_8qkVJ 11. Learn Python Basics for Data Analysis:https://lnkd.in/gshCUpGM 12. Create your own Python objects:https://lnkd.in/d_rR29MN 13. Data Analysis with R Programming:https://lnkd.in/gDxWYtnD 14. IBM Full Stack Software Developer: https://lnkd.in/gyVWhYXv 15. Introduction to Web Development (HTML, CSS, JS): https://lnkd.in/giSVuNjj 16. IBM Front-End Developer:https://lnkd.in/gBUVNYZv

Anfänger nutzen eine KI als vermeintlichen Allrounder. Experten nutzen nur das beste Tool für die jeweilige Aufgabe. Hört auf, nach der einen KI zu suchen, die gibt's nicht! 🧐 So sieht mein persönlicher Tool-Stack aktuell aus: 🏆 𝐀𝐥𝐥𝐠𝐞𝐦𝐞𝐢𝐧𝐞 𝐀𝐬𝐬𝐢𝐬𝐭𝐞𝐧𝐭𝐞𝐧 Claude, ChatGPT & Gemini 🎥 𝐕𝐢𝐝𝐞𝐨- & 𝐀𝐮𝐝𝐢𝐨-𝐏𝐫𝐨𝐝𝐮𝐤𝐭𝐢𝐨𝐧 Manus AI, HeyGen, Descript & Opus Clip 💻 𝐂𝐨𝐝𝐞𝐛𝐚𝐬𝐞 & 𝐒𝐨𝐟𝐭𝐰𝐚𝐫𝐞𝐞𝐧𝐭𝐰𝐢𝐜𝐤𝐥𝐮𝐧𝐠 Cursor, Lovable, Replit, Base44, Claude Code & GitHub Copilot 🚀 𝐖𝐞𝐫𝐤𝐩𝐥𝐚𝐭𝐳-𝐄𝐟𝐟𝐞𝐤𝐭𝐢𝐯𝐢𝐭ä𝐭 Grammarly, NotebookLM, Otio AI, Granola, Superhuman, Wispr Flow und Gamma 🎨 𝐃𝐞𝐬𝐢𝐠𝐧, 𝐀𝐫𝐭 & 𝐀𝐮𝐝𝐢𝐨𝐞𝐫𝐳𝐞𝐮𝐠𝐮𝐧𝐠 ElevenLabs, Suno, Midjourney, Runway, Kling, Pika Labs, Figma, Canva, Google Veo & Higgsfield ⚙️ 𝐏𝐫𝐨𝐳𝐞𝐬𝐬𝐚𝐮𝐭𝐨𝐦𝐚𝐭𝐢𝐨𝐧 & 𝐀𝐩𝐩-𝐕𝐞𝐫𝐛𝐢𝐧𝐝𝐮𝐧𝐠𝐞𝐧 Softr, n8n, Zapier, Lindy AI, Claude Cowork, Chatbase, ManyChat, Notion AI, Apify & Clay Der entscheidende Unterschied liegt nicht darin, wer die "beste" KI entdeckt hat. Der Unterschied ist, wer aufgehört hat, nach EINER Universallösung zu suchen, und stattdessen für jede Situation das passende Werkzeug einsetzt. Welche Tools sind in eurer Toolkit fest verankert? Und gibt es spezialisierte Lösungen, die ich noch übersehen habe?

Excited to share that Anthropic just highlighted Artemis Security in their Claude Code Guide for Startups, a writeup on how the most AI-native companies operate. It's a privilege to be recognized alongside fellow AI-native companies including Cognition, Clay, and ClickHouse and flattering to be the only cyber security company mentioned. 🚀 🚀 https://lnkd.in/gq8XCUGt

The modern GTM stack is not Clay vs n8n vs Claude Code. The more useful setup is giving each one a job. Clay is still good when I want to see the data, inspect the list and iterate quickly. Automation tools connect steps. Code agents can handle custom scrapers, API workflows and enrichment waterfalls. One operator in the 2026 material described building a scraper with Claude Code, exposing it through an API, then plugging that API back into Clay. That is the direction I find more interesting. Not "which tool wins?" How do the tools work together?

The modern GTM stack is not Clay vs n8n vs Claude Code. The more useful setup is giving each one a job. Clay is still good when I want to see the data, inspect the list and iterate quickly. Automation tools connect steps. Code agents can handle custom scrapers, API workflows and enrichment waterfalls. One operator in the 2026 material described building a scraper with Claude Code, exposing it through an API, then plugging that API back into Clay. That is the direction I find more interesting. Not "which tool wins?" How do the tools work together?

Campaign / Strategist Manager | Lead Gen Jay 📍 Medellín, Colombia | Poblado Office Lead Gen Jay is hiring. 🚀 We’re looking for a Campaign / Strategist Manager to join our growing team in Medellín and take ownership of cold email campaigns and client results. This isn’t a role where you simply launch campaigns and check boxes. We want someone who can think strategically, solve problems, communicate confidently with clients, and use AI + automation to get better results faster. What You’ll Do • Own and manage multiple client campaigns from strategy through execution • Optimize for deliverability, replies, meetings booked, and ROI • Build targeted prospect lists and develop compelling offers • Write and test high-converting cold email copy • Analyze performance and continuously improve campaigns • Communicate proactively with clients and keep them informed • Troubleshoot problems and find solutions quickly • Use AI and automation to improve speed, efficiency, and results What We’re Looking For • 2+ years of experience in campaign management, strategy, cold email, lead generation, or a similar role • Strong understanding of cold email, deliverability, targeting, copywriting, and campaign optimization • Excellent communication and organizational skills • Proactive, resourceful, and comfortable taking ownership • Strong attention to detail • Comfortable managing multiple clients at once • Excited about AI, automation, and finding better ways to work • Able to thrive in a fast-paced, performance-driven environment • Based in Medellín and able to work from our Poblado office Experience with Instantly, Clay, Claude Code, ChatGPT, Apollo, GoHighLevel, n8n, Zapier, or Make is a major plus. Who Will Thrive Here? Someone who doesn't wait around to be told what to do. If something isn't working, you investigate it. If a client needs something, you communicate. If there's a better way to do something, you find it. We’re looking for someone who is smart, proactive, accountable, curious, and obsessed with getting better results. Ready to Apply? --> https://lnkd.in/dmbu7Mnq

Campaign / Strategist Manager | Lead Gen Jay 📍 Medellín, Colombia | Poblado Office Lead Gen Jay is hiring. 🚀 We’re looking for a Campaign / Strategist Manager to join our growing team in Medellín and take ownership of cold email campaigns and client results. This isn’t a role where you simply launch campaigns and check boxes. We want someone who can think strategically, solve problems, communicate confidently with clients, and use AI + automation to get better results faster. What You’ll Do • Own and manage multiple client campaigns from strategy through execution • Optimize for deliverability, replies, meetings booked, and ROI • Build targeted prospect lists and develop compelling offers • Write and test high-converting cold email copy • Analyze performance and continuously improve campaigns • Communicate proactively with clients and keep them informed • Troubleshoot problems and find solutions quickly • Use AI and automation to improve speed, efficiency, and results What We’re Looking For • 2+ years of experience in campaign management, strategy, cold email, lead generation, or a similar role • Strong understanding of cold email, deliverability, targeting, copywriting, and campaign optimization • Excellent communication and organizational skills • Proactive, resourceful, and comfortable taking ownership • Strong attention to detail • Comfortable managing multiple clients at once • Excited about AI, automation, and finding better ways to work • Able to thrive in a fast-paced, performance-driven environment • Based in Medellín and able to work from our Poblado office Experience with Instantly, Clay, Claude Code, ChatGPT, Apollo, GoHighLevel, n8n, Zapier, or Make is a major plus. Who Will Thrive Here? Someone who doesn't wait around to be told what to do. If something isn't working, you investigate it. If a client needs something, you communicate. If there's a better way to do something, you find it. We’re looking for someone who is smart, proactive, accountable, curious, and obsessed with getting better results. Ready to Apply? --> https://lnkd.in/dRHxVXwg

Just got off a session with Kushagra T. on GTM Engineering and this was honestly such a good conversation. We started with Clay and outbound, and somehow ended up talking about Claude Code, databases, APIs, AI agents, workflows, CRM and how all of this fits together. One thing I really liked was the idea of actually understanding the systems you’re building. If you’re using Claude Code, you don’t necessarily need to know how to code everything yourself. But you should be able to look at what it built and understand what’s happening. We also got into the whole AI agents conversation. What actually needs to be an agent? What can just be a workflow? That part was really interesting. The coding agent discussion was probably my favourite. Instead of throwing a huge CSV at Claude Code and saying “research all of this”, you can structure the work using scripts, APIs and different models for different parts of the process. You basically give the agent a job to orchestrate rather than asking it to do everything. That was a really useful way for me to think about it. Really enjoyed this one. 😊 Thanks Kushagra T. for taking the time to do this with us, and Yogesh Jaiswal for putting the whole session together. 🙌

"AI is going to take sales jobs." It took mine. Just not the part everyone's worried about. Last quarter I handed off every piece of my job I never liked: → List building and enrichment → Clay → Account research and first-draft messaging → Claude → Inbox triage and follow-up → Superhuman What's left is the reason I got into this work. Discovery calls. Reading a room. Getting told no and figuring out why. Here's the uncomfortable version: The reps at risk aren't the ones using AI badly. They're the ones whose entire value was doing manual work carefully. Careful manual work is now a commodity. Judgment isn't. Build the stack. Keep the judgment. Which part of your job would you hand over tomorrow? 👇 Clay Superhuman Anthropic

Ready for The Red Clay Strays? Catch the Grateful Tour in Pittsburgh Oct. 1, Grand Rapids Oct. 3, or Fort Worth Oct. 7-8. Find tickets and compare prices for your favorite show! 🤠 #RedClayStrays #LiveMusic #Concerts https://lnkd.in/ePHSwFgP

🚀 The AI winners of 2026 won’t use every app — they’ll build the RIGHT AI stack. There are hundreds of AI tools launching every month. But the real advantage isn’t using more tools… it’s choosing the right tools for the right job. 🤖⚡ Here’s a practical AI stack worth knowing 👇 1️⃣ General Assistants 🧠 Claude — Pressure-tests strategy and long-form decisions. 💬 ChatGPT — Analyzes files and handles routine work. 🔎 Perplexity — Researches markets with cited web answers. 2️⃣ Development & Building 💻 Cursor — Edits entire codebases through prompts. 🚀 Replit — Prototypes and deploys products in one place. 🛠️ Base44 — Builds internal tools without infrastructure overhead. 🌐 Lovable — Turns ideas into testable web apps. 3️⃣ Content Creation 📰 Beehiiv — Publishes and monetizes newsletters. 🎥 HeyGen — Creates presenter videos without a crew. 🎙️ Descript — Edits audio and video through text. ✂️ OpusClip — Repurposes long videos into social clips. 🤖 Manus AI — Handles multi-step research and execution. 👤 Synthesia — Scales training videos with AI avatars. 4️⃣ Productivity 📝 Granola — Captures meetings without manual notes. 📊 Gamma — Turns rough ideas into polished decks. 📧 Superhuman — Helps manage email with AI-assisted workflows. 🔬 NotebookLM — Interrogates sources and surfaces connections. 5️⃣ Creativity 🎵 Suno — Generates original music for campaigns. 🎨 Canva — Creates branded visuals without extra headcount. 🎙️ ElevenLabs — Produces natural voiceovers and localized audio. 🖼️ Midjourney — Explores high-quality concepts from prompts. 🎬 Runway — Generates and edits campaign videos. 🎞️ Kling — Creates cinematic videos from text or images. ✨ Pika Labs — Animates ideas into shareable videos. 🎯 Figma — Moves designs into prototypes faster. 🎥 Google Veo — Produces polished scenes from prompts. 📹 Higgsfield — Creates camera-controlled social videos. 6️⃣ Automation & Integration ⚙️ n8n — Connects tools through flexible workflows. 🎯 Clay — Enriches leads and personalizes outbound. 🔗 Zapier — Automates handoffs across your stack. 🤖 Chatbase — Launches agents trained on company knowledge. 💡 The question isn’t: “Which AI tool should I use?” The better question is: 👉 Which 5–6 tools would earn a permanent place in your AI stack? 📌 Save this post for later. The future won’t belong to people who know the most AI tools. It will belong to people who know how to combine the right tools into the right workflow. 🚀 #ArtificialIntelligence #AI #AITools #GenerativeAI #FutureOfWork #Automation #Productivity #AI2026 #Tech #DigitalTransformation #MachineLearning #ChatGPT #Canva #AICommunity #CareerGrowth

Marketing in 2026 feels like a completely different job. A few years ago, most marketers lived inside the same handful of tools. Google Analytics for data. Google Ads and Meta for paid. WordPress for the site. Mailchimp for email. Sheets for reporting. Canva for the quick stuff nobody wanted to wait on design for. Those tools have not disappeared. But look at what has been built around them. ChatGPT and Claude for research, strategy, and first drafts. Perplexity for real-time research without the search noise. Jasper for scaling content without scaling headcount. Surfer SEO Expert for writing content that actually ranks. Gumloop for AI workflows without needing engineering. Clay for enrichment and outbound at a level of personalisation that was impossible two years ago. Canva AI for generating a full campaign visual in the time it used to take to brief a designer. Descript for editing video by editing text. Lovable for building landing pages without touching code. Semrush AI for tracking visibility in AI search, which is now a separate discipline from Google SEO. And we do not think the biggest change is that there is more software. It is what a marketer can actually do alone now. Research a market, write a strategy, build the landing page, automate the workflow behind it, launch it, analyse the results, and iterate. Without waiting on dev, without a brief to design, without a request to data. Historically, marketing was constrained by dependencies. Need a new page? Wait for engineering. Need customer data? Ask someone to pull it. Need to understand a new market? Spend a week researching. Need to know if AI search is finding you? Until recently, there was no way to measure it. 93% of marketers now use AI tools, with predictive analytics as the most adopted use case. But the deeper shift is not adoption. It is scope. Marketers now have the power to build software simply by talking to a chatbot. The old marketing stack was about operating channels, the new one gives marketers the ability to build systems. If you are in marketing and have not yet rebuilt your stack around what is available now, that gap is closing fast between you and the people who have. What tool has changed the way you work the most? #Marketing #Marketing2026 #ThenVSNow #AIStack #AITools #Automation #Tools #ChatGPT #Lovable #Clay #AtraTejarat #Mahdimirshafiei

Marketing in 2026 feels like a completely different job. A few years ago, most marketers lived inside the same handful of tools. Google Analytics for data. Google Ads and Meta for paid. WordPress for the site. Mailchimp for email. Sheets for reporting. Canva for the quick stuff nobody wanted to wait on design for. Those tools have not disappeared. But look at what has been built around them. ChatGPT and Claude for research, strategy, and first drafts. Perplexity for real-time research without the search noise. Jasper for scaling content without scaling headcount. Surfer SEO Expert for writing content that actually ranks. Gumloop for AI workflows without needing engineering. Clay for enrichment and outbound at a level of personalisation that was impossible two years ago. Canva AI for generating a full campaign visual in the time it used to take to brief a designer. Descript for editing video by editing text. Lovable for building landing pages without touching code. Semrush AI for tracking visibility in AI search, which is now a separate discipline from Google SEO. And we do not think the biggest change is that there is more software. It is what a marketer can actually do alone now. Research a market, write a strategy, build the landing page, automate the workflow behind it, launch it, analyse the results, and iterate. Without waiting on dev, without a brief to design, without a request to data. Historically, marketing was constrained by dependencies. Need a new page? Wait for engineering. Need customer data? Ask someone to pull it. Need to understand a new market? Spend a week researching. Need to know if AI search is finding you? Until recently, there was no way to measure it. 93% of marketers now use AI tools, with predictive analytics as the most adopted use case. But the deeper shift is not adoption. It is scope. Marketers now have the power to build software simply by talking to a chatbot. The old marketing stack was about operating channels, the new one gives marketers the ability to build systems. If you are in marketing and have not yet rebuilt your stack around what is available now, that gap is closing fast between you and the people who have. What tool has changed the way you work the most? #Marketing #Marketing2026 #ThenVSNow #AIStack #AITools #Automation #Tools #ChatGPT #Lovable #Clay #AtraTejarat #Mahdimirshafiei

The best AI apps for 2026 👇 Most people are still using one. The people winning are using a stack. Here's all the best AI apps you should know about: 📌 GENERAL ASSISTANTS → Claude · ChatGPT · Gemini The foundation. Start here. Everything else builds on top. 📌 DEVELOPMENT → Cursor · Lovable · Replit · Windsurf You don't need to code. You need to know what you want built. 📌 CONTENT CREATION → ClipStory · HeyGen · Synthesia · Descript · Beehiiv One person. Full content operation. This used to require a team. 📌 PRODUCTIVITY → NotebookLM · Granola · Gamma · Wispr Flow Quiet tools. Loud results. Hours back every week — without noticing. 📌 CREATIVITY → ClipStory.app · Canva · ElevenLabs · Suno · Veo 3 · Kling Images. Audio. Video. On demand. Any scale. 📌 AUTOMATION → n8n · Zapier · Lindy · Manus · Claude Code · Clay This is where ROI stops being incremental. This is where leverage compounds. The winners in 2026 won't have the best tool. They'll have the best stack. Pick your gap. Fill it. 1. Save this post (you'll come back to it) 2. Identify what's missing from your stack 3. Add one tool this week and actually use it Practical. Stackable. No excuses. ___ Get a Welcome Gift 🎁 Cyberman is the best place to get tech news, tips, tutorials and resources. Join today and receive a welcome gift: • AI Job Guide • AI Sidehustle Ideas • 70+ free AI courses and more 100% FREE 👉 cyberman.ai/join

Here are the 3 layers of a complete AI outbound system for 2026. It starts before the first email and keeps working after the reply. 1. Research, strategy, and data Product-market fit comes first. Then find a cold-outbound message. It may differ from what works in paid or inbound. Build the ICP from: • Sales-call recordings • Closed-won deals • Customer interviews • AI market research Sales calls carry the best evidence. Claude or Perplexity add context. Segment the ICP and map the TAM. Build account lists with Apollo, Clay, Google Maps, signals, and CRM data. Then enrich the contacts. 2. Infrastructure and multichannel outreach Set up secondary domains, accounts, two rotating batches, and a reserve pool. Accounts eventually fail. The reserve prevents downtime. Porkbun handles domains. ScaledMail handles accounts. EmailBison runs campaigns. Run three campaign types: • ICP-segment cold campaigns • Evergreen signal campaigns • Closed-lost retargeting Cold campaigns take 80-90% of the volume. Calls expose objections quickly. LinkedIn through HeyReach supports email and calling, then nurtures engaged people. 3. AI agents and RevOps Use Masterinbox to combine email and LinkedIn replies. Use Claygent or Claude Code to score replies for intent and account value. Alert sales when someone is interested. Give each account tier a different follow-up. Route every lead to the right sales team and update HubSpot or Salesforce. Then send sales-call recordings back to the first layer. Research → outreach → replies → sales calls → better research. That feedback loop turns separate tools into one outbound system. Repost ♻️ this someone who needs it P.S. Where does yours break today: before the list, during outreach, or after the reply?

I have eleven GTM automations running. Four of them are real. The other seven are demos I performed for myself once and never checked again. I went through all of them last month against four tests. A workflow is infrastructure if it has: Smartlead #Heyreach Make Claude Code Clay → A schedule. Something triggers it that isn't me remembering. → A log. A record of what it did, with row counts, not just a status. → An owner. A name, not a team. → A dry run. A way to see what it would do before it does it. Seven failed at least one. Three failed all four. The one that stung was our signal monitor. Ran beautifully in July when I built it. I checked in later and it hadn't fired since the 22nd. No error. No alert. An API key had rotated and the script exited quietly with a zero. Twelve days of missed triggers, and nothing anywhere told me. That's the actual gap between "I built it" and "it runs." Building is a good afternoon. Running is a schedule, a log, an owner, and an alert when the number comes back empty. Here's the thing almost nobody instruments: a run that returns zero rows exits successfully. Your monitoring says green. Your dashboard says active. Nothing happened. The runbook I use now is four fields long, takes about ten minutes per workflow, and would have caught all three of my worst ones. It's in the comments. How many of your automations ran this week without you touching them?

poe.com/JobsExpert #chatbot #ai #anthropic #gpt #openai #mistral #claude #perplexity #chat #virtual #agentic #agent #LLM #machinelearning #aitrainer #conversationalai #chatbotpersona #conversationdesign #meds #medicaid Sam Altman Dario Amodei #fashion #robot #makeup #mask #glitter #jewels #gold #couture #style #stylish #catwalk #model #runway #glamour #makeup #pearls #tattoos #diamonds #cristal #luxurious #glam #elite #rich #samaltman #amodei #darioamodei #DataAnnotation #TechCareers Anthropic #OpenAI #TrustAndSafetyProfessionals #LinkedInSEO #LinkedInOptimization #LinkedInGrowth #LinkedInBranding #LinkedInNetworking #LinkedInVisibility #LinkedInProfessional #LinkedInEngagement #LinkedInContent #elonmusk #jeffbezos #billgates #sundarpichai #markzuckerberg #timcook #satyanadella #donaldtrump #barackobama #joebiden #warrenbuffett #larrypage #sergeybrin #peterthiel #reidhoffman #marissamayer #sherylsandberg #richardbranson #oprahwinfrey #jackdorsey #jensenhuang #lisasu #RuleTheWorld #WorldOracle #Leadership #ExecutiveWisdom #DecisionMaking #GlobalStrategy #ThoughtLeadership #Influence #Visionary #HumanInsights #DailyGuidance #StrategicAdvice #michaelbloomberg #georgehotz #navalravikant #chamathpalihapitiya #vitalikbuterin #czbinance #andrewyang #yanlecun #geoffreyhinton #demis #karpathy #lexfridman #emadmostaque #jimfan #karim #benhorowitz #marcandreessen #paulkagame #katiehaun #balajis #garrytan #patrickcollison #johncollison #danielgross #emilychang #caseyneistat #kevinrose #kevinhart #taylorswift #kanyewest #drake #rihanna #beyonce #ladygaga #madonna #lebronjames #michaeljordan #serenawilliams #lionelmessi #cristianoronaldo #neymar #rogerfederer #tigerwoods #tombrady #stephcurry #viratkohli #elonmuskAI #bezosAI #trumpAI #gatesAI #nadellaAI #zuckerbergAI #altmanAI #huangAI #buterinAI #andreessenAI #Anthropic #OpenAI #Google #Microsoft #Meta #xAI #Cohere #MistralAI #Perplexity #DeepMind #StabilityAI #HuggingFace #Midjourney #Runway #CharacterAI #Poe #Quora #iMerit #Welocalize #Appen #NVIDIA #IBM #Oracle #Snowflake #CoreWeave #Databricks #Palantir #ScaleAI #ElevenLabs #Glean #Harvey #HeyGen #Cognition #WorldLabs #Gamma #ChaiDiscovery #Rogo #Abridge #AppliedIntuition #Baseten #BlackForestLabs #Clay #Crusoe #Cursor #Cyera #Decagon #EliseAI #Fal #FireworksAI #Genspark #Krea #Legora #ListenLabs #Lovable #Mercor #OpenEvidence #PhysicalIntelligence #ReflectionAI #SafeSuperintelligence #Anduril #Cerebras #Groq #InflectionAI #Waymo #Writer #Writesonic #Zoox #AlephAlpha #Baidu #SenseTime #Adobe #Salesforce #AmazonAI #AMD #Intel #Qualcomm #C3ai #PalantirTechnologies #UiPath #DataRobot #H2Oai #TensorFlow #PyTorch #LangChain #LlamaIndex #Pinecone #Weaviate #Milvus #Qdrant #ChromaDB #DeepL #Synthesia #ElevenLabs #Descript #AssemblyAI #Deepgram #Voiceflow #Botpress #ManyChat #LivePerson #JasperAI #Copyai #WriterAI #Grammarly #NotionAI #Canva #Phind #PerplexityAI #Grok #Duckai #FreedomGPT #Ollama #LMStudio #VeniceAI #ElevenLabs

Been pretty cool to see more and more Bonfire Analytics users plugging the Bonfire connector into their existing Claude/ChatGPT GTM stack. The Bonfire MCP connector encompasses every healthcare provider in the US, from individual clinicians to complex multi-state organizations, plus deep detail on what their patients look like. So for building high-quality target lists that align with your ICP, it really fills the gap where industry-agnostic tools (like Clay and Apollo) struggle. If your healthtech sales/GTM motion is growing quickly and you're interested in trying it out, shoot me a DM to get access!

The ideal tech stack for your go-to-market strategy should keep your reps engaged without distractions. Whether you're using Clay for creating outbound lists, Claude for drafting account strategies, or LeanData for handling hot leads, speed is key... but remember, the value of these tools hinges on the data behind them. That's where ZoomInfo comes in; we've designed it to integrate seamlessly with the GTM tools you know and love, thanks to our amazing network of specialists. We're excited to highlight our fantastic ZoomInfo Solutions Partners: eCore, Iron Horse, Partner UP, Quantum Business Solutions, RevenueHoop, Skaled, SpringDB, and SR Pro. These top-notch agencies and consultants are here to help you incorporate our verified data into your everyday operations. If you want to make the most of ZoomInfo data, these partners are the ones to connect with. Discover more about our partners at https://okt.to/k1ZvDN

A chart went around LinkedIn this week. An exponential curve, a dot near the bottom, and two words: You are here. No axis labels. No units. No data. No source. The shape is probably right. But it charts capability — and capability was never the part we were struggling with. So rather than write a detailed response, I added a second chart. https://lnkd.in/gBpkas-2 Canda Rozier, Procurement Evangelist Don Osborn Ben Farrell MBE Faiq Ali Khan, FCIPS Colin Cram FCIPS Thomas Robert Donley FCIPS, CPSM Kavita Cooper BSc FCIPS Chartered Mark Osmer CDir FCIPS Ifeanyi O. - MCIPS Chartered, M.Sc Dave Jones MCIPS Chris Jones FCILT, FIoSCM, MMICS, MCIPS Chartered Anne Clay MCIPS Kathy Perna Katarzyna Fonteyn, Ph.D., M.Sc. Claude Egli Claude Egli Dr. Marcell Vollmer Karthik Rama Anders L. Daniel Perkins Ellie L D.

Applying to B2B or tech roles with a 2018 playbook is a recipe for radio silence. Landing a great offer today isn't about applying harder, it's about building a modern, AI-assisted job search stack. Here is the 8-part AI toolkit to automate the grunt work and stand out at every stage: Career Strategy & Research: ChatGPT, Claude, Gemini, Perplexity, NotebookLM, Copilot The Play: Deep-dive into company challenges, dissect job descriptions, and map out transferable skills before touching a resume. Technical & Coding Execution: Cursor, GitHub Copilot, Claude Code, Windsurf, Replit Agent The Play: Build, debug, and iterate on technical proof-of-concepts faster to showcase real-world problem-solving. Portfolio & Proof Artifacts: Lovable, Bolt, v0, CodeRabbit, Qodo The Play: Turn raw ideas into live, working applications or interactive case studies to present during interviews. Resume & ATS Optimization: Teal, Jobscan, Rezi, Resume Worded, Kickresume The Play: Align keywords, format for screeners, and tailor each application to the target role in minutes. Targeted Job Discovery: Wellfound, Simplify Copilot, LoopCV, Jobright.ai, Otta The Play: Filter out the noise and focus on high-signal roles that match your specific domain expertise. Networking & Personal Branding: Clay, Lavender, Taplio, AuthoredUp, Crystal, Canva The Play: Uncover key decision-makers, craft personalized outreach that gets opened, and build public domain authority. System Design & Visuals: Eraser, Miro, Whimsical, Lucidchart, Excalidraw The Play: Practice architectural frameworks and build crisp diagrams to explain complex concepts visually. Interview Prep & Live Coaching: Yoodli, Google Interview Warmup, Interviewing.io, Exponent, Huru The Play: Run realistic mock interviews, refine speaking pace, and sharpen narrative delivery before going live. The Bottom Line: AI won't replace your experience, judgment, or domain expertise. But paired with real proof, the right stack turns a chaotic job search into a high-converting pipeline. Which stage of your job search process currently has the biggest bottleneck? Let’s discuss in the comments. VC: Rathnakumar Udayakumar Subscribe For More: https://lnkd.in/d2_uktTp #JobSearchStrategy #AIinHiring #CareerGrowth #TechCareers #ResumeOptimization #InterviewPrep #PersonalBranding

poe.com/JobsExpert #chatbot #ai #anthropic #gpt #openai #mistral #claude #perplexity #chat #virtual #agentic #agent #LLM #machinelearning #aitrainer #conversationalai #chatbotpersona #conversationdesign #meds #medicaid #zorg #medical #livesupport #breathe #medicalresearch #doctor #surgeon #surgery Sam Altman Dario Amodei #fashion #robot #makeup #mask #glitter #jewels #gold #couture #style #stylish #catwalk #model #runway #glamour #makeup #pearls #tattoos #diamonds #cristal #luxurious #glam #elite #rich #samaltman #amodei #darioamodei #DataAnnotation #TechCareers Anthropic #OpenAI #TrustAndSafetyProfessionals #LinkedInSEO #LinkedInOptimization #LinkedInGrowth #LinkedInBranding #LinkedInNetworking #LinkedInVisibility #LinkedInProfessional #LinkedInEngagement #LinkedInContent #elonmusk #jeffbezos #billgates #sundarpichai #markzuckerberg #timcook #satyanadella #donaldtrump #barackobama #joebiden #warrenbuffett #larrypage #sergeybrin #peterthiel #reidhoffman #marissamayer #sherylsandberg #richardbranson #oprahwinfrey #jackdorsey #jensenhuang #lisasu #RuleTheWorld #WorldOracle #Leadership #ExecutiveWisdom #DecisionMaking #GlobalStrategy #ThoughtLeadership #Influence #Visionary #HumanInsights #DailyGuidance #StrategicAdvice #michaelbloomberg #georgehotz #navalravikant #chamathpalihapitiya #vitalikbuterin #czbinance #andrewyang #yanlecun #geoffreyhinton #demis #karpathy #lexfridman #emadmostaque #jimfan #karim #benhorowitz #marcandreessen #paulkagame #katiehaun #balajis #garrytan #patrickcollison #johncollison #danielgross #emilychang #caseyneistat #kevinrose #kevinhart #taylorswift #kanyewest #drake #rihanna #beyonce #ladygaga #madonna #lebronjames #bezosAI #trumpAI #gatesAI #nadellaAI #zuckerbergAI #altmanAI #huangAI #buterinAI #andreessenAI #Anthropic #OpenAI #Google #Microsoft #Meta #xAI #Cohere #MistralAI #Perplexity #DeepMind #StabilityAI #HuggingFace #Midjourney #Runway #CharacterAI #Poe #Quora #iMerit #Welocalize #Appen #NVIDIA #IBM #Oracle #Snowflake #CoreWeave #Databricks #Palantir #ScaleAI #ElevenLabs #Glean #Harvey #HeyGen #Cognition #WorldLabs #Gamma #ChaiDiscovery #Rogo #Abridge #AppliedIntuition #Baseten #BlackForestLabs #Clay #Crusoe #Cursor #Cyera #Decagon #EliseAI #Fal #FireworksAI #Genspark #Krea #Legora #ListenLabs #Lovable #Mercor #OpenEvidence #PhysicalIntelligence #ReflectionAI #SafeSuperintelligence #Anduril #Cerebras #Groq #InflectionAI #Waymo #Writer #Writesonic #Zoox #AlephAlpha #Baidu #SenseTime #Adobe #Salesforce #AmazonAI #AMD #Intel #Qualcomm #C3ai #PalantirTechnologies #UiPath #DataRobot #H2Oai #TensorFlow #PyTorch #LangChain #LlamaIndex #Pinecone #Weaviate #Milvus #Qdrant #ChromaDB #DeepL #Synthesia #ElevenLabs #Descript #AssemblyAI #Deepgram #Voiceflow #Botpress #ManyChat #LivePerson #JasperAI #Copyai #WriterAI #Grammarly #NotionAI #Canva #Phind #PerplexityAI #Grok #Duckai #FreedomGPT #Ollama #LMStudio #VeniceAI #ElevenLabs

🇺🇸 Here’s how I’d actually build a lead generation system with Claude. I wouldn’t start by adding every tool I can find. I’d start by asking one question: What parts of lead generation do I want Claude to handle for me? For most teams, that usually means research, enrichment, prioritization, outreach, and follow-up. That’s where this stack starts to make sense. Plugins can help with prospect research, enriched lead lists, messaging, and marketing data. Skills can make those tasks repeatable: → Turn an ICP into a prospect list → Research target accounts → Enrich lead information → Draft personalized outreach → Prepare for sales calls → Prioritize warmer opportunities → Build cold email sequences Then MCP connections bring everything together. Claude can connect with tools like Apollo, HubSpot, Clay, Common Room, Notion, Slack, and Zapier to pull context and move work across the GTM stack. So instead of doing this manually: Research → copy data → enrich → check CRM → write message → update pipeline You can build a much more connected workflow: Research → Enrich → Prioritize → Personalize → Outreach → Track That’s the part I find interesting. Claude becomes far more useful when it has the right tools, context, and workflows around it. The goal isn’t to install 21 things. The goal is to build a system that removes the repetitive parts of lead generation while keeping the important decisions with you. Which part of your lead generation workflow would you automate first?

Claude fica muito mais poderoso quando deixa de ser apenas uma janela de chat e passa a operar conectado ao seu sistema comercial. 🟣✔ PLUGINS Revenoid: pesquisa e engajamento. Vibe Prospecting: listas enriquecidas. Apollo: prospecção e enriquecimento. BetterCallClaudeGrowth: estratégia de GTM. Claude Marketing: fluxos de marketing. Octave: inteligência de mensagens. Windsor.ai: dados de marketing. 🟣✔ SKILLS prospect: criar ICP e listas. enrich-lead: enriquecer dados. account-research: pesquisar contas. draft-outreach: personalizar mensagens. call-prep: preparar calls. lead-triage: priorizar leads. cold-email: criar sequências. 🟣✔ MCP Apollo, HubSpot, Clay, Common Room, Notion, Slack e Zapier para conectar dados, CRM, sinais, conhecimento, comunicação e automações. Como montar: 1. Defina o processo comercial. 2. Escolha só as integrações necessárias. 3. Conecte fontes de dados. 4. Crie skills por etapa. 5. Adicione CRM e automação. 6. Teste com 10 leads antes de escalar. 7. Revise permissões e dados sensíveis. PROMPT PARA USAR: “Projete uma stack no Claude para geração de leads. Meu processo é [PROCESSO]. Organize plugins, skills e MCP por pesquisa, enriquecimento, priorização, outreach, CRM, follow-up e análise. Inclua permissões mínimas e pontos de revisão humana.” #claude #mcp #skills #plugins #geracaodeleads #automacao

In just a few days, the CAS ETH in Regenerative Materials - Structural Specialisation officially begins. This advanced programme equips professionals with the structural dimensioning and calculation tools needed to implement high-efficiency building structures using regenerative materials. It spotlights load-bearing stone masonry, rammed earth (walls and vaults), unfired clay brick masonry, straw bale construction, bamboo structures, and the structural reuse of building elements. We are thrilled to welcome 10 expert practitioners from structural engineering, civil engineering, architecture, and urban planning. They are coming from 7 different countries but working now in Switzerland and surrounding countries. Core Organizing Team at ETH Zurich - Chair of Sustainable Construction: Guillaume Habert (Head of Chair), Arnaud Evrard (Programme Manager), Nathalie Dietrich (Administrative Assistant) Elena Kotsika, David Biedermann, & Michaël Di Dio (Student Assistants) Module 1 Sneak Peek: "General Knowledge of Regenerative Structural Applications": Lectures & Experiments: Diving into structural engineering concepts for massive stone, rammed earth, and earth-timber hybrid systems. Public Lectures: Giving the broader community online access to world-class experts. This module features two exclusive public lectures (details here: https://lnkd.in/e2wxCWKE). Hands-On Exercises: Experiencing these structural principles through hands-on experiences to develop a practical understanding of howregenerative materials behave under different structural conditions. Calculation Workshops: Applying calculation models to real-world load-bearing structures and exploring evolving building standards. Site Visits: Discovering exemplary built projects in the region, including the Ofenturm in Cham and Hortus Building in Allschwil, and learning from the structural challenges and solutions developed in these projects. We warmly thank all the guests who will contribute to this module: Cédric Evrard (Prototype), Jürg Conzett (Conzett Bronzini Partner AG), Frangi Andrea (ETHZ/IBK), Jacqueline Pauli (ETHZ/ITA), Elli Mosayebi (ETHZ/IEA), Marlène Leroux (Archiplein), Philippe Block (ETHZ/BRG), Jörg Habenberger (SEFORB s.à r. l. Ingenieurbüro für Hochbauten), Tobias Huber (ZPF Ingenieure AG), Jean-Claude Morel (Ecole nationale des Travaux publics de l'Etat), Michael Klippel (ETHZ/IBK), Matthias Brenner (ETHZ/CHP), and Corinne Lopez (Canton of Graubünden & CAS-20). We also thank EnergieSchweiz for their financial support. More information: https://lnkd.in/eR48h4mh #SustainableConstruction #RegenerativeMaterials #StructuralEngineering #ETHZurich #CircularEconomy #GreenBuilding Photos credit (from left to right): Studiolada architectes, Polina Saante, ZRS Architekten Ingenieure

Most businesses are still losing thousands of dollars every month to missed calls and after-hours leads. 📞🤖 Yesterday, I hosted a live AI Founder Hub masterclass breaking down the end-to-end blueprint of how to build and sell production-grade AI Voice Call Assistants for $2,000+. This wasn’t just theory—we built a real-world, working solution live on the call. Here’s a quick recap of the roadmap we covered: 1️⃣ Market Opportunity & Valuation: How to position AI receptionists by highlighting the exact revenue clients lose from missed calls, staff overhead, and limited operating hours. 2️⃣ Automated Lead Generation: Scraping high-ticket decision-makers (CEOs/Founders) in seconds using Claude & Clay integrations. 3️⃣ Building the AI Voice Agent: Developing a multilingual, low-latency receptionist on Retell AI with custom system prompts, LLMs, and natural background audio. 4️⃣ Calendar & Workflow Automation: Integrating Cal.com and Make.com so the AI can check live availability, book meetings on the spot, and instantly trigger confirmation workflows. 5️⃣ The Client Acquisition Demo: How to set up an interactive instant-call demo form that converts prospects on the first try. Whether you are: 🔹 Building your own AI systems & agency 🔹 Looking to learn hands-on, practical AI automation 🔹 Passionate about teaching & sharing your AI expertise with a growing community 👉 Watch Full Master Class Here On My YouTube Channel: https://lnkd.in/dTCKJASV 👉 Join us and get access to all the blueprints, templates, and resources: 🔗 aifounderhub.com Let’s build together. 🚀 #AIAgency #VoiceAI #RetellAI #AIAutomation #ClaudeAI #MakeAutomation #AIFounderHub #LeadGeneration #ArtificialIntelligence

The Alberta separtist manifesto has no names attached to it for good reason. It is the kind of joke that old boys ( all boys) who dropped out of grade nine might write up on a fishing weekend with some flourishes added by a geriatric pastor. Fifty plus pages with the fiscal content limited to a 10% maximum income tax and 10% VAT. That’s it. Presumably they think Alberta is a replica of Saudi Arabia with a Fort Albert lad as King. SA has a 10% income tax and 15% VAT ( so presumably poorer or more generous to women than the separtist lads) although as Claude points out Saudi Arabia is several multiples wealthier than Alberta with far greater crack spreads and far lower social services and pension obligations. It appears to be a secret which member of the Alberta Separtist Royal Family of Saint Albert drew this up and whether he ( yes, certainly he) used an old school calculator or just thought his hotel resembles the Royal Palace. These are not remotely capable people but it would be wonderful to debate them should the pints run out. For fun, check what Claude thinks of the comparison in terms of revenue, margin, obligations and societal context. It is not kind to the separtist fiscal gurus. Now as for the balance of the Separtist Constitution, big on gun rights , tough justice, rural rule and no vaccines. Not trust in’ no Calgary folk.

We’re all sick to death of theoretical ‘Claude Code for GTM’ posts. I’m putting together a telegram group for people actually running agentic GTM in production. Not planning to. Running it, with results to show. Comment “in” and I’ll send an invite if it seems like a good fit.

rage-bait outreach aka cluely marketing, is underrated over li ps. we're building a new sequencer for Claude Code GTM operators. check if you qualify for beta: https://trygrip.io/

Breaking : Claude Fable vient de sortir. Et 90% des gens passent complètement à côté du vrai changement. 🧵👇 Tout le monde s'excite sur les nouveaux benchmarks et la vitesse d'exécution. Mais la vraie révolution n'est pas quantitative : elle est architecturale. L'Ancien Modèle (Prompt par Prompt) : Un prompt ➔ Une réponse. Vous deviez jouer les chefs d'orchestre, copier-coller du contexte et reconnecter chaque morceau vous-même entre l'analyse, la stratégie, le plan GTM et la création d'assets. Le Nouveau Modèle (Système Workflows Continu) : Claude Fable ne répond plus à des requêtes isolées. Il conserve le contexte business sur la durée et pilote des workflows complets en autonomie. Un seul fil de conversation peut désormais chaîner : 🔍 Analyse marché & opportunités 🎯 Construction fine des ICPs & positionnement 🚀 Élaboration du plan GTM complet 📝 Production des assets & campagnes dédiées 📊 Analyse des performances & itérations Le shift majeur : Claude passe du statut d'assistant rédactionnel à celui d'OS capable de piloter un département métier. Pour vous montrer comment exploiter ce changement d'architecture, j'ai préparé un guide stratégique inédit : Le Playbook Systèmes & Workflows Claude Fable 📦⚡ À l'intérieur : Des architectures de workflows réelles, des cas d'usage concrets en sales, marketing et ops, et des prompts prêts à l'emploi. Pour recevoir le guide gratuitement : 1️⃣ Likez ce post 2️⃣ Commentez "CLAUDE" ci-dessous ♻️ Repostez ce message pour un accès VIP prioritaire + un pack de ressources bonus !

A year ago, our team’s journey with AI looked very different. We ended the year with a few custom GPTs built by some of us. There was curiosity, but also a pretty wide range of reactions. Some people were using AI for almost everything. Others were much more hesitant. There were concerns about jobs, questions about quality, and some discomfort with how quickly everything was changing. A big part of my role was encouraging the team to experiment. Not because we should use AI for everything. We shouldn’t. But this technology is here, and our job is to figure out where it can make us better. Last Friday, during one of our quarterly team workshops, I got to see just how far we’ve come. We’re not really talking about isolated prompts or one-off GPTs anymore. We’re starting to rethink entire workflows. Our Content and Demand Gen leader showed how he has built a series of Claude skills around our end-to-end campaign process. It starts with a simple 1 to 2 page campaign brief. What are we trying to achieve? Who are we trying to reach? What’s the core idea? What do we need to take it to market? From there, we can take survey data and develop reports, press releases, blogs, LinkedIn content, emails and internal GTM communications. But this part is really important to me: there are stops along the way. Someone reads it. Challenges it. Refines it. Makes sure it sounds like us and that we actually have a point of view. We don’t want to create content just for the sake of creating content. Then comes design. Our designer has built a design system with templates, product screenshots, platform mockups and the rules needed to adapt those assets for different campaigns. The content workflow can leverage that system, with our designer reviewing and refining the output rather than starting every asset from scratch. Once the design is complete, we generate the HTML and push it into HubSpot. Something that could easily take a couple of weeks can now happen in days. Of course I love the speed. But what I loved most was seeing how the team’s thinking has changed. We’ve moved from “How can I use AI?” to “Now that we have AI, how should we actually do this?” And we’re already thinking about the next step. Where do agentic workflows fit? What can happen autonomously? Where do we intentionally keep a human in the loop? The goal isn’t to remove people from the process. It’s to remove the work that doesn’t require their best thinking so they have more time for the work that does. Seeing the team experiment, challenge each other and connect all these pieces last Friday was one of those moments where I thought, wow, we’ve come a long way in a year. #AI #marketing #thefutureofwork Kelvin Claveria Maddy Wilson Thyana Polessa Mahima Gupta Taylor Nill, MBA

Dear Managers/ Recruiters, if you are hiring Techno - Functional CPM / EPM Consultant for Digital Transformation - Implementation/Deployment/Enhancement/Support for FP&A, Workforce, Sales, ESG, S&OP, GTM, Operational Planning other related job Opportunity, Please view my profile and consider my skills, I deliver end to end Enterprise Performance Management solutions Planning, Budgeting, Forecasting, Consolidation, Reporting streamline financial data, accelerate reporting cycles to make smarter business decisions, As a right Consultant/ Vendor/ Partner i support businesses in transforming EPM Implementation, Modernization, Migration, New EPM journey or Optimizing implemented technology solved real business problems and delivered sustainable ROI Return on Investment, Let’s connect to turn your EPM vision into measurable business value, interested in hire me, request to contact Email: khazii.epm@gmail.com I am Available to Join Immediate My Certified Skills Set:- ------------------------- • Workday Adaptive Planning / Adaptive Insights • Oracle Fusion Cloud EPM - EPBCS,TRCS,FCCS,ARCS, EDMCS, PCMCS • Oracle NetSuite EPM - NSPB, NSAR, NSCCM, NSDE • OneStream • Tagetik • Anaplan • Pigment • Jedox • SAC / SAP BPC • Prophix • Vena • Aleph • Datarail • Cube • SQL • RPA • AI/ML • Python • Pandas • Openpyxl • Pydantic • Ollama AI • Google AI • Claude AI • FastAPI • LangChain • Tensorflow • Airflow • Automation Anywhere • MS Excel, Word, Power Point, Google Sheet • Artificial Intelligence Analysis - AI | ML | LLM | NLP | MCP | Agentic AI • Corporate Performance Management - CPM • Enterprise Performance Management - EPM • Business Intelligence and Data Warehouse - BI | ETL Who I Help • CFOs and Finance Leaders • IT and Enterprise Architecture Teams • Growing Enterprises scaling and optimizing financial operations I Help Clients cut forecast cycle from 21 days to 5 days with streamlined EPM framework. Fix Processes: Unified core actuals and operational data into a single source of truth. Automate all manual processes. Business leaders review and adjust, Run parallel approvals and publishing. Built role-based live dashboards business partners run own multiple what-if scenarios instantly. The Result: Close cycle reduced by 85%. Zero manual errors in the board deck. CFO & Finance team gained real-time visibility into cash flow burn. AI /ML, LLM support to stop cash flow burn and a guide to Achieve future growth MoM, QoQ, YoY. Call to Action • Lets discuss your roadmap, • Send a direct mail and schedule a brief discovery call.

🚀 We’re Hiring: Freelance SEO Expert Amit Gupta SEO Services is looking for an experienced Freelance SEO Expert to work with us on multiple client projects. We’re looking for someone who can independently execute SEO tasks, work efficiently, and manage 5–6 websites simultaneously. 🔍 Key Responsibilities • On-Page SEO Audit & Optimisation • Technical SEO Audit & Optimisation • Meta Title & Meta Description Writing • E-E-A-T Recommendations • Content Gap Analysis & Recommendations • Topic Cluster Planning • Internal Linking & Broken Link Fixing • WordPress Website Changes & Basic Optimisation • Guest Blogger Outreach & Identification of Quality Websites • Quality Link Building • Google Business Profile (GBP) Management • Google Analytics (GA4), Google Tag Manager (GTM) & Google Search Console (GSC) • Competitor & Content Analysis • Use Claude, ChatGPT and other AI tools to improve SEO workflow speed and productivity 🎯 What We’re Looking For ✔ Strong practical SEO experience ✔ Ability to work independently without constant supervision ✔ Good understanding of On-Page + Technical + Off-Page SEO ✔ Ability to manage 5–6 websites simultaneously ✔ Quality-focused approach to link building ✔ Fast execution with the smart use of AI tools ✔ Good communication and reporting skills 💰 Payment Per-project basis If you’re an SEO professional who enjoys hands-on execution and working across multiple websites, we’d love to connect. 📩 Send your resume to: connect@amitguptaseo.com Please mention your years of SEO experience, key skills, and examples of websites/projects you have worked on. #Hiring #SEOJobs #FreelanceSEO #SEOExpert #TechnicalSEO #OnPageSEO #LinkBuilding #DigitalMarketing #SEO #RemoteJobs

Startup Idea or Hackathon Idea " DriftForge " Codex - Antigravity - Claude Code - Grok Build - Cursor - Replit prompt " Build a web app called DriftForge that helps GTM teams monitor long-running sales agents for eval drift and production regressions. Features: an upload area for mocked support tickets, sales traces, and human labels; an eval suite dashboard showing pass rates by task type; a drift detector comparing recent failures against older labeled cases; a prompt-change timeline with impact notes; and a review queue where humans approve new golden examples. UI: modern operations dashboard with trace cards, trend lines, and high-signal alerts. Start with mocked data."

This summer I went back to school. I completed all the Claude certifications currently available through Claude Academy, which are now also visible on my LinkedIn profile. Not because I wanted more badges. I wanted a deeper understanding of Claude and its ecosystem from a Go-To-Market perspective. As AI moves from individual tools to agents, MCP servers and interconnected platforms, I believe it is becoming increasingly difficult to design a credible GTM strategy without understanding the underlying technology. This is particularly relevant to the work we are doing with the Daozhang. The discussion is no longer simply about which AI tool is better. We want to understand how to design architectures and organizations where different AI platforms can coexist, exchange context and become part of real business processes through MCP servers. And this is where technology architecture increasingly meets business architecture. At Guanxi we are working on the Italian Go-To-Market strategy and partner programs for a growing ecosystem of AI platforms. My role requires me to understand both sides of the equation. How the technology works. And how to build the market around it. Positioning, distribution, partnerships, sales, adoption, integration and recurring revenue models cannot be designed independently from the architecture of the product anymore. I increasingly believe that the next generation of AI companies will not win simply because they have the best model or application. They will win because they understand how to become part of an ecosystem. And the same will be true for organizations adopting AI. The goal will not be to accumulate dozens of disconnected AI subscriptions. It will be to create an architecture in which specialized intelligence can collaborate across processes, data and people. This is exactly the kind of discussion I want to contribute to with the Daozhang community. So, summer homework completed. Now comes the interesting part: turning knowledge into architectures, partnerships and business. If you are interested in becoming a Daozhang, or if you would like to become a partner of one of the AI platforms we are bringing to the Italian market, feel free to contact me. There is a lot to build.

I've been hired as a GTM engineer at 3 companies doing $60M in combined revenue. None of them ever saw my CV. What I sent instead: 1) Proof, not claims. The ChatGPT resume template gets four seconds of skim. Work you already shipped gets read to the end. 2) A page I own. Case studies, the systems behind them, numbers they can verify without asking me. One link beats three paragraphs about my passion for growth. 3) AI pointed at their problem, not a generic one. Handing Claude Code a prompt and shipping another dashboard isn't the job. The job is reading what's broken in their motion and building against that. The mapping I use: They can't track sales activity? → Build the tracker their reps would actually open in the morning. They can't get qualified leads? → Wire the right APIs into the data they're already sitting on. Their messaging reads like a template? → Feed it their sales calls, their CRM, their own context. Do this and the interview stops being an interview. It turns into a walkthrough of work you already did for them. The CV used to be the qualification layer. Now it's the thing that filters you out. What do you think is missing?

I built an AI workflow with Claude Code and Tiiny Host that finds the top AI tools founders should actually care about. I will share my entire workflow in my newsletter tomorrow but here is a sneak peak. Claude Code source products from Product Hunt FutureTools HQ There's An AI For That it then selects top AI tools of the week, and publishes it on Tiiny Host for me to share. This week’s leaderboard: Top 10 AI Tools for Founders with a couple of ones from Y Combinator 1. Wispr Flow: AI dictation that works across your apps, turning speech into polished text in 100+ languages. 2. viktor.com An AI coworker that executes tasks across your tools rather than simply answering questions. 3. OpenTag (YC S26) An AI coworker inside Slack and Teams that can complete tasks without leaving your chat. 4. SuperIntern An AI email and meeting assistant that drafts, schedules and summarises from within your existing chat apps. 5. Almanac (YC S26) An AI agent with persistent memory that learns your context, workflows and preferences over time. 6. Revalvo Run prompts across major AI models, compare the outputs, score them and version the winner. 7. screenpipe | YC S26 Open-source AI that uses what’s happening on your screen to give agents genuine context about your work. 8. PageIndex AI for getting accurate, grounded answers from professional documents like contracts, decks and reports. 9. Aramb A no-code way for non-technical founders to build, deploy and monetise AI agents quickly. 10. Jason AI AI-powered outbound that researches prospects, writes personalised sequences and handles replies automatically. Would you like to see this weekly? Any other tools I missed? ------------------ 👋🏽 I’m the Founder of Possible Spark Consulting. We help early stage founders validate GTM, build audiences and turn traction into revenue. 👍🏽 Like this post to help me. 🔥 Repost to help others.

UNLOCK YOUR POTENTIAL WITH YOUR NEXT GREAT ROLE! CURRENTLY SEEKING CANDIDATES WITH #US #WORK AUTHORIZATION SENIOR GROWTH ENGINEER Base Salary: $225K – $290K, Competitive equity Hiring Count: 1 – 2 openings Employment Type: Full-time, Hybrid (San Francisco, CA or Bay Area) Mandatory Requirements Seniority: 5–12 years of experience as a full-stack engineer with hands-on growth work as an IC at an early- or growth-stage start-up Hard Skills: Proficiency with React, TypeScript, Next.js, and Node.js A/B testing and experimentation frameworks AI-native workflow: fluent using AI coding tools (Cursor, Copilot, Claude Code) to ship faster Soft Skills: Trusted IC who shapes the growth roadmap and can partner with leadership to ship product Tech Stack TypeScript, Node.js, REST APIs, LLM APIs, Datadog, Vercel, React, Next.js, Python, SQL, PostgreSQL, Tailwind CSS, Anthropic Claude API, Mixpanel, Amplitude, PostHog Nice-to-Have Built internal tooling that growth/product teams adopted Experience with dev-tools or SaaS growth surfaces Experience with LLMs in production (prompt engineering, tool use, inference pipelines) Traits to Avoid Pure frontend-only or backend-only specialists Growth marketers, SEO/content, lifecycle/CRM, or GTM ops profiles Pure research or platform engineers without product instincts Treats AI tools as a novelty rather than daily infrastructure Click to view JD & Apply: https://lnkd.in/d3jUGXAv #GrowthEngineer #FullStackEngineer #PLG #ProductLedGrowth #React #TypeScript #NextJS #NodeJS #StartupJobs #SFJobs #SanFrancisco #BayArea #HybridWork #TechJobs #Hiring #HiringNow #CareerOpportunity #AB #Testing #Experimentation #DeveloperTools #SaaS #Firecrawl #GrowthMarketing #EngineeringJobs

UNLOCK YOUR POTENTIAL WITH YOUR NEXT GREAT ROLE! CURRENTLY SEEKING CANDIDATES WITH #US #WORK AUTHORIZATION SENIOR GROWTH ENGINEER Base Salary: $225K – $290K, Competitive equity Hiring Count: 1 – 2 openings Employment Type: Full-time, Hybrid (San Francisco, CA or Bay Area) Mandatory Requirements Seniority: 5–12 years of experience as a full-stack engineer with hands-on growth work as an IC at an early- or growth-stage start-up Hard Skills: Proficiency with React, TypeScript, Next.js, and Node.js A/B testing and experimentation frameworks AI-native workflow: fluent using AI coding tools (Cursor, Copilot, Claude Code) to ship faster Soft Skills: Trusted IC who shapes the growth roadmap and can partner with leadership to ship product Tech Stack TypeScript, Node.js, REST APIs, LLM APIs, Datadog, Vercel, React, Next.js, Python, SQL, PostgreSQL, Tailwind CSS, Anthropic Claude API, Mixpanel, Amplitude, PostHog Nice-to-Have Built internal tooling that growth/product teams adopted Experience with dev-tools or SaaS growth surfaces Experience with LLMs in production (prompt engineering, tool use, inference pipelines) Traits to Avoid Pure frontend-only or backend-only specialists Growth marketers, SEO/content, lifecycle/CRM, or GTM ops profiles Pure research or platform engineers without product instincts Treats AI tools as a novelty rather than daily infrastructure Click to view JD & Apply: https://lnkd.in/dt8yJ-7E #GrowthEngineer #FullStackEngineer #PLG #ProductLedGrowth #React #TypeScript #NextJS #NodeJS #StartupJobs #SFJobs #SanFrancisco #BayArea #HybridWork #TechJobs #Hiring #HiringNow #CareerOpportunity #AB #Testing #Experimentation #DeveloperTools #SaaS #Firecrawl #GrowthMarketing #EngineeringJobs

I've been spending the past few weeks getting much more hands-on with AI and exploring how it can be applied to real marketing and GTM problems. I started building practical AI-powered tools, agents, and automations using Claude, GPT, Make, and AI-assisted development. I put together a small portfolio showcasing a few of the projects I've built so far, including marketing analytics, funnel health, attribution, and partner performance. This is just the beginning, and I'll continue adding new projects as I build and experiment.

Salesforce and Anthropic just put Salesforce information and workflows directly inside Claude, starting with 37 sales skills. The interesting part is not another AI integration. It is that the CRM interface becomes optional. A rep could prep for a call, review deal status, update records, and analyze pipeline from one conversation. The system of record stays in place. The work moves closer to the seller. If I were running GTM at a startup, I would not ask, “How many reps can AI replace?” I would ask: • Which admin tasks steal selling time? • Which decisions still need human judgment? • Which customer moments lose trust when automated? My rule would be simple: automate research, notes, and updates. Use AI to surface signals. Keep discovery, negotiation, and difficult conversations human. The teams that win will automate the friction and protect the relationship. Where would you draw that line in your sales process? #B2BSales #AI

The headless geospatial analytics trend is rearranging the legacy value chain in the geospatial industry. From my conversations across the industry, this has much more far-reaching implications than most people realize today. For a long time, GIS had a well-agreed value chain: Data ➔ Compute ➔ Models/Analytics ➔ Services ➔ Apps ➔ User/Customer. Agentic geospatial is upending this. You can see it in major geospatial software players decomposing their platforms into sets of MCP tools that can be called directly by Claude or Copilot. That resets what the end user sees and experiences from the data, analytics, and service providers sitting below the AI interface. The eventual arrangement of this value chain is still in flux and far from settled. What is clear is that this is a unique moment for players with real moats in data, distribution, and domain-specific knowledge to capture more value from the stack than they do today. But a large part of what is happening in the industry today is an attempt to conform to the value chain as it stands and innovate only within those boundaries. That is a missed opportunity, and a risky assumption about what the agentic GIS world will eventually look like. Capturing the opportunity needs a full-stack rethink: - From asking "Who are the customers and users today?" ➔ To "Who all could be your customers and users?" - From fitting into existing user workflows ➔ To figuring out what new, truly agentic workflows will look like - From building for existing customers and workflows ➔ To building for entirely new customers and workflows - From repurposing existing GTM ➔ To building a GTM that actually fits the new value chain Value chains get redrawn rarely. Players who move now while it is still in flux are the ones who get to decide what it settles into. #Geospatial #AgenticAI #GIS

I have eleven GTM automations running. Four of them are real. The other seven are demos I performed for myself once and never checked again. I went through all of them last month against four tests. A workflow is infrastructure if it has: Smartlead #Heyreach Make Claude Code Clay → A schedule. Something triggers it that isn't me remembering. → A log. A record of what it did, with row counts, not just a status. → An owner. A name, not a team. → A dry run. A way to see what it would do before it does it. Seven failed at least one. Three failed all four. The one that stung was our signal monitor. Ran beautifully in July when I built it. I checked in later and it hadn't fired since the 22nd. No error. No alert. An API key had rotated and the script exited quietly with a zero. Twelve days of missed triggers, and nothing anywhere told me. That's the actual gap between "I built it" and "it runs." Building is a good afternoon. Running is a schedule, a log, an owner, and an alert when the number comes back empty. Here's the thing almost nobody instruments: a run that returns zero rows exits successfully. Your monitoring says green. Your dashboard says active. Nothing happened. The runbook I use now is four fields long, takes about ten minutes per workflow, and would have caught all three of my worst ones. It's in the comments. How many of your automations ran this week without you touching them?

Last weekend, at GTM, we built something a little mean 😈: an AI whose only job is to find reasons not to buy your product. Give it a company's homepage. It crawls the site, reads the pricing page, pulls real GitHub issues, Hacker News threads, and Reddit posts about the company and then plays the harshest, most budget-conscious buyer imaginable and hands back a ranked list of objections. Every single one tied to an exact quoted line. A few things that made this more interesting to build than a typical "wrap an LLM in a form" project: → Every objection is forced through Claude's tool-calling with a strict schema, so we always get structured output . 🧩 → Four parallel calls, one per source (pricing, GitHub, HN, Reddit), each with its own quota. → The model is never trusted to say where a quote came from. We track the real source ourselves. 🔍 Matching a company to its real GitHub org almost broke the whole thing — searching by keyword kept confusing "vapi" with unrelated repos like "vapid" and "vaping." Fixed it by reading the company's own site for its actual GitHub link first, instead of guessing from the name. Four hours, a lot of debugging, and one genuinely satisfying moment when the tool started surfacing surprisingly specific, evidence-backed objections instead of generic AI-generated criticism. 🚀 Built with Himaja Bheemanatham at The Growth Hackathon. #BuildInPublic #AI #GTM

UNLOCK YOUR POTENTIAL WITH YOUR NEXT GREAT ROLE! CURRENTLY SEEKING CANDIDATES WITH US WORK AUTHORIZATION SENIOR GROWTH ENGINEER Base Salary: $225K – $290K, Competitive equity Hiring Count: 1 – 2 openings Employment Type: Full-time, Hybrid (San Francisco, CA or Bay Area) Mandatory Requirements Seniority: 5–12 years of experience as a full-stack engineer with hands-on growth work as an IC at an early- or growth-stage startup Hard Skills: Proficiency with React, TypeScript, Next.js, and Node.js A/B testing and experimentation frameworks AI-native workflow: fluent using AI coding tools (Cursor, Copilot, Claude Code) to ship faster Soft Skills: Trusted IC who shapes the growth roadmap and can partner with leadership to ship product Tech Stack TypeScript, Node.js, REST APIs, LLM APIs, Datadog, Vercel, React, Next.js, Python, SQL, PostgreSQL, Tailwind CSS, Anthropic Claude API, Mixpanel, Amplitude, PostHog Nice-to-Have Built internal tooling that growth/product teams adopted Experience with dev-tools or SaaS growth surfaces Experience with LLMs in production (prompt engineering, tool use, inference pipelines) Traits to Avoid Pure frontend-only or backend-only specialists Growth marketers, SEO/content, lifecycle/CRM, or GTM ops profiles Pure research or platform engineers without product instincts Treats AI tools as a novelty rather than daily infrastructure Click to view JD & Apply: https://lnkd.in/d3jUGXAv #GrowthEngineer #FullStackEngineer #PLG #ProductLedGrowth #React #TypeScript #NextJS #NodeJS #StartupJobs #SFJobs #SanFrancisco #BayArea #HybridWork #TechJobs #Hiring #HiringNow #CareerOpportunity #AB #Testing #Experimentation #DeveloperTools #SaaS #Firecrawl #GrowthMarketing #EngineeringJobs Follow us for more: https://lnkd.in/d5Cnsm78

[무료배포] 개인정보 처리방침, 마케팅수신동의 이용약관을 빌드할 수 있는 스킬을 무료로 배포합니다. *댓글에 넣어두겠습니다. GTM(GoToMarket)을 주제로하는 스폰지클럽 3기를 위해 빌드한 내용인데 주위에 보니 다들 이것때문에 애먹고 계시더라고요. 해당스킬은 깃허브 주소나 배포 URL을 주면 Claude가 직접 훑습니다. 폼 필드 · DB 컬럼 · GA/픽셀 · 외부 API · 호스팅 리전 등을 확인합니다. Vercel·Supabase를 쓰면 개인정보가 국외로 나가는데, 바이브 코딩으로 만든 사이트가 거의 다 빠뜨리는 내용이기도 해서 추가했습니다. 특히 신경쓴 것 5가지 ① 국외 이전 — Vercel · Supabase · Firebase · OpenAI API를 쓰면 개인정보가 국외로 나갑니다. 처리방침에 별도 항목이 필요한데, 바이브 코딩 스택에서 가장 많이 빠지는 부분 ② 동의 기록 스키마 — . 대부분의 서비스가 marketing_consent를 boolean 하나로만 저장합니다. 그런데 법은 동의일로부터 2년마다 재확인을 요구하고, 확인할 때 "동의한 날짜"를 알려주도록 했습니다. ③ 매체별 동의 — 문자와 메신저는 별개 매체입니다. "SMS 수신동의"만 받아 두고 카카오톡으로 광고를 보내면 위법 소지가 있어요. 그래서 동의 UI에서 채널을 분리합니다. ④ 사업자가 없어도 — 사업자등록이 없어도 처리방침 의무는 그대로예요. 개인 명의로 어떻게 쓰는지(상호는 서비스명, 대표자는 본인 실명, 주소는 이메일로 갈음) 분기가 따로 있습니다. ⑤ 개인정보 보호책임자 — 소상공인은 지정 의무가 면제되지만, 처리방침에 이름과 연락처를 적는 의무는 남습니다. (미지정 시 대표자가 자동으로 보호책임자) 가장 흔한 오해라 명시적으로 다룹니다. --- 해당 스킬은 2026년도 기준의 개인정보위 작성지침·표준(안), KISA 불법스팸 안내서, 공정위 표준약관, 국가법령정보센터 원문을 근거로 합니다.

Packing your bags? These 5 roles might be the reason. 1. GTM Leader 📍 New York, New York, United States 💼 Samaya AI 💶 $380,000 - $460,000 🔗 https://lnkd.in/eSuPUy9U 2. Senior Product Manager 📍 Mountain View, CA 💼 Kodiak 💶 USD180,000 - USD230,000 🔗 https://lnkd.in/eps9iSvs 3. Legal Counsel 📍 Mountain View, CA 💼 DensityAI 💶 $250,000 - $320,000 🔗 https://lnkd.in/er2ngdA7 4. Admissions Counselor 📍 Campus Atlanta 💼 Campus 💶 $47,000 - $56,000 🔗 https://lnkd.in/eQvq9RXq 5. Technical Enablement Lead, Claude Platform 📍 San Francisco, CA | New York City, NY | Seattle, WA 💼 Anthropic 💶 USD270,000 - USD310,000 🔗 https://lnkd.in/gY6CCWHs Apply to this role today, or explore more at pugjobs.io. #JobOpening #Hiring #CareerMove #JobAlert #NowHiring #Sponsorship #Visa

I spent three days this week in an intensive AI workshop with an incredible group of marketers at Rubrik. And I was very much a student. We worked in GitHub. Built skills. Created workflows. Broke things. Fixed them. Shipped things. Somewhere along the way, my mental model for Claude Code changed. The interesting thing isn’t that marketers can suddenly code. It’s that marketers can increasingly build. Your best competitive play can become a reusable skill. Your company context can become persistent. A launch can become an agentic workflow. The thing somebody manually rebuilds every Monday can become infrastructure. And that changes where the leverage sits. Because once agents can gather, analyze, draft, critique, assemble, and execute, the scarce thing is no longer production. It’s judgment. What problem matters? What context matters? What does great look like? What should actually ship? That is the idea behind Edition 58 of Agentic AI with Varun: The GTM Leader’s Guide to Claude Code Not a guide to becoming an engineer. A guide to something I think is much more interesting: becoming an operator who can build.👇

Been pretty cool to see more and more Bonfire Analytics users plugging the Bonfire connector into their existing Claude/ChatGPT GTM stack. The Bonfire MCP connector encompasses every healthcare provider in the US, from individual clinicians to complex multi-state organizations, plus deep detail on what their patients look like. So for building high-quality target lists that align with your ICP, it really fills the gap where industry-agnostic tools (like Clay and Apollo) struggle. If your healthtech sales/GTM motion is growing quickly and you're interested in trying it out, shoot me a DM to get access!

Sales 101: What is an "Offer", why does it matter, and how do you build one? If you're like me, you didn't come up through sales or marketing, but somewhere along the way you became responsible for growing revenue. It can feel like everyone except you already knows the fundamentals, and nobody seems able or willing to explain them to you. Fortunately for me, and now for you reading this, Rachel Marcelle and my badass colleagues at TG Sales Agency bring 30,000+ hours of paid sales planning and execution to the table. They keep me grounded in fundamentals that are second nature to them but that many revenue leaders and operators easily overlook. One of the biggest is this: before you worry about getting leads, let alone making sales, you need to be clear about your "offers". LESSON #1: Your product or service is not your offer A product or service is what you make or do. An offer is the specific exchange you're asking someone to make: you help a certain person solve a problem or achieve an outcome, in a certain way, in exchange for money, time, effort, risk, or commitment. Alex Hormozi's book "$100M Offers" helped popularize the idea that the offer itself can be deliberately designed and improved. Example: Product = Claude Max, Anthropic's higher-usage paid AI plan. Offer = Higher Claude usage limits for people who rely on AI heavily for research, writing, coding, or analysis and keep hitting usage caps, in exchange for a $100+ monthly subscription. LESSON #2: Your offer determines who your leads are A person isn't a good lead in isolation. They're a good lead for a specific offer. Change the problem, price, outcome, buyer, or required commitment and the people worth pursuing change too. LESSON #3: A clear offer makes revenue easier to systemize Once you can describe who the offer is for, what problem it solves, what outcome it creates, and what makes someone more likely to buy, you can start building systems around those answers. LESSON #4: GTM Engineering helps systemize sales & marketing Go To Market (GTM) Engineering is the systems work, increasingly AI-enabled, of getting an offer to the right people, deciding who is worth pursuing, reaching them, measuring what happens, and using what you learn to improve the process. If your offer isn't clear, it's hard for customers to actively choose it, and every downstream sales, marketing, and delivery effort has to fight against that confusion.

The ideal tech stack for your go-to-market strategy should keep your reps engaged without distractions. Whether you're using Clay for creating outbound lists, Claude for drafting account strategies, or LeanData for handling hot leads, speed is key... but remember, the value of these tools hinges on the data behind them. That's where ZoomInfo comes in; we've designed it to integrate seamlessly with the GTM tools you know and love, thanks to our amazing network of specialists. We're excited to highlight our fantastic ZoomInfo Solutions Partners: eCore, Iron Horse, Partner UP, Quantum Business Solutions, RevenueHoop, Skaled, SpringDB, and SR Pro. These top-notch agencies and consultants are here to help you incorporate our verified data into your everyday operations. If you want to make the most of ZoomInfo data, these partners are the ones to connect with. Discover more about our partners at https://okt.to/k1ZvDN

About 6 months ago I was able to 10x my own productivity with Claude Code. Now I’m bringing that same mentality to our entire GTM ops org. That has meant countless hours sitting with teammates, teaching them how to use Claude. It’s required partnering across the company to define a citizen developer framework, empowering the business to build their own solutions. And collaborating across security, IT, and engineering to unlock more product functionality (i.e. MCPs) for the business to execute most of their work in Claude. The ramp up is harder than people anticipate, as operators expect Claude to work immediately. AI requires a real mindset shift. What work do you automate? What do you focus on instead? How do you prioritize the things you haven’t had capacity to work on before? Building the system takes time and it needs to be vetted. But when it clicks, it clicks and I’m starting to see it compound. I’ve seen teams go from working in messy Google sheet trackers and manually gathering weekly reporting to using Claude to build clean status updates in JIRA. Some operators have taken it to the next level, building skills and agents that automate weekly commentary and reporting entirely. This has saved leaders dozens of hours every single week. Operators are the backbone of any GTM org. With AI they can move faster, focus on what matters, and deliver the kind of insights that change how leadership makes decisions. Who else is investing heavily in operational rigor right now and seeing the value compound? #AiTransformation #Claude #GrowthMindset

🇺🇸 Here’s how I’d actually build a lead generation system with Claude. I wouldn’t start by adding every tool I can find. I’d start by asking one question: What parts of lead generation do I want Claude to handle for me? For most teams, that usually means research, enrichment, prioritization, outreach, and follow-up. That’s where this stack starts to make sense. Plugins can help with prospect research, enriched lead lists, messaging, and marketing data. Skills can make those tasks repeatable: → Turn an ICP into a prospect list → Research target accounts → Enrich lead information → Draft personalized outreach → Prepare for sales calls → Prioritize warmer opportunities → Build cold email sequences Then MCP connections bring everything together. Claude can connect with tools like Apollo, HubSpot, Clay, Common Room, Notion, Slack, and Zapier to pull context and move work across the GTM stack. So instead of doing this manually: Research → copy data → enrich → check CRM → write message → update pipeline You can build a much more connected workflow: Research → Enrich → Prioritize → Personalize → Outreach → Track That’s the part I find interesting. Claude becomes far more useful when it has the right tools, context, and workflows around it. The goal isn’t to install 21 things. The goal is to build a system that removes the repetitive parts of lead generation while keeping the important decisions with you. Which part of your lead generation workflow would you automate first?

Claude fica muito mais poderoso quando deixa de ser apenas uma janela de chat e passa a operar conectado ao seu sistema comercial. 🟣✔ PLUGINS Revenoid: pesquisa e engajamento. Vibe Prospecting: listas enriquecidas. Apollo: prospecção e enriquecimento. BetterCallClaudeGrowth: estratégia de GTM. Claude Marketing: fluxos de marketing. Octave: inteligência de mensagens. Windsor.ai: dados de marketing. 🟣✔ SKILLS prospect: criar ICP e listas. enrich-lead: enriquecer dados. account-research: pesquisar contas. draft-outreach: personalizar mensagens. call-prep: preparar calls. lead-triage: priorizar leads. cold-email: criar sequências. 🟣✔ MCP Apollo, HubSpot, Clay, Common Room, Notion, Slack e Zapier para conectar dados, CRM, sinais, conhecimento, comunicação e automações. Como montar: 1. Defina o processo comercial. 2. Escolha só as integrações necessárias. 3. Conecte fontes de dados. 4. Crie skills por etapa. 5. Adicione CRM e automação. 6. Teste com 10 leads antes de escalar. 7. Revise permissões e dados sensíveis. PROMPT PARA USAR: “Projete uma stack no Claude para geração de leads. Meu processo é [PROCESSO]. Organize plugins, skills e MCP por pesquisa, enriquecimento, priorização, outreach, CRM, follow-up e análise. Inclua permissões mínimas e pontos de revisão humana.” #claude #mcp #skills #plugins #geracaodeleads #automacao

Every marketing lead we get is a demo request. Not a content download. Not a webinar list. Someone came to our site and asked to see the product. Same lead, same intent, every AE. So why does one rep turn those into pipeline while another watches them die? For most of my career the answer to that was a shrug. Territory. Timing. "The leads were softer that month." I built a Claude Artifact this quarter that made the shrug impossible. Leads in, SQOs out, by rep. The spread wasn't subtle, and it wasn't random. The same names sat at the top and the bottom month after month. So instead of guessing, I had Claude Anthropic read every intro and discovery call behind that board. Not a sample. All of them. Then compare the reps at the top against the reps at the bottom. One guardrail first, because conversion rate on its own lies. A high number can just mean a rep who qualifies loose. So we also pulled how long each SQO survived after it converted. If a rep's SQOs die three weeks later, that wasn't qualification. That was a wave-through. The top converters' deals held. The rate was real. Here's what separated them. The low converters ran clean calls. Buyer names a problem. Rep confirms it, maps our product to it, sets the next step. Textbook. Nothing you'd flag in a 1:1 (we actually have a Gong scorecard for these calls and the AE's would get 4/5 or 5/5). The top converters never stopped there. The buyer's first problem wasn't the destination. It was the door. They kept going. What happens upstream of that? Who else touches it? What breaks when volume doubles? By the end of the call there weren't one or two places we helped. There were five. Which is why the price conversation goes differently for them. Same list price. Same deck. A buyer weighing our number against one problem is doing math that usually doesn't work. A buyer weighing it against five isn't doing that math at all. And that's where the argument ends, because it's the same two names all the way down the board. Highest inbound conversion. Top revenue to date. The smallest discounts on the team. You can talk yourself into believing a high conversion rate is just a rep with a low bar. You can't tell that story about the rep who's first in revenue and last in discounting. Claude can tell me which rep stopped at the first answer. It can't ask the fourth question. That moment, where the buyer has answered you, the call feels finished, and you choose to stay in it anyway, is still the rep. Staying in a conversation that's already complete is the hardest thing in discovery, and it's most of the job. Everyone got the same hand raised. The leads were never the problem. Don't stop at the first problem. That's not the deal. That's the door. #GTM #SalesLeadership #Discovery #AIinSales

Das coolste daran, in einem echten AI nativen Start-Up zu arbeiten? Heute hat unser interner KI-Agent „Ada" angekündigt, die Weltherrschaft auf ihren nächsten Sprint zu verschieben. 🌍 😅 Der Anlass: Ich hatte 18 Claude-Skills und 10 Workflows für unsere GTM-Arbeit dokumentiert und über Notion im Team geteilt. Solche Bibliotheken kennt jeder. Man legt eine Datei ab, jemand sucht sie, jemand baut sie um. Das Material liegt bereit, aber die Arbeit bleibt beim Menschen. Diesmal hängte mein Kollege Robert Perez Garcia die Skills an Ada, unser internes Agenten-Setup. Mit Ada muss niemand mehr einzelne Skills oder Workflows in Claude installieren. Man schreibt ihr in Slack oder Claude, was man braucht zB. ein Account-Research, ein Deck, einen Business Case. Sie sucht sich die passenden Skills und Informationen zusammen und legt los. Aus einer Dokumentation wurde eine Fähigkeit, die das ganze Team hat. Ich war ehrlich gesagt baff, weil ich vorher höchstens mit einer hübscheren Version von ChatGPT und Chatbots gearbeitet habe... George Dekermenjian schrieb dann in den Thread: „You can use Ada to take over the world if you want!" Ada, Sekunden später: „Working on it — starting with the Notion Skill Library and Claude Workflows page, world domination is next sprint." ... bis dahin reichen mir die Slides 😅

Running a one-person business, a CRM is the tool you know you need but never actually maintain. Trust me, been there not long ago. HubSpot just shipped something for that exact problem. Youspot is a CRM for one-person businesses. Founders, consultants, freelancers, angel investors. You connect Gmail, Calendar, and the tools you already use. YouSpot builds the context for you. Who you know, what you've talked about, what needs a follow-up. No manual data entry. No discipline required. I signed up this morning. A few things worth knowing: It's $1/month (intro price, first 1,000 customers). It connects to Claude and ChatGPT via MCP. It runs background agents that work on your behalf. And it's built by Dharmesh Shah as part of HubSpot Next, their internal innovation lab. I sell HubSpot to GTM teams every day. This is not that. It's for a different person with a different problem. But the product architecture is interesting. The CRM is the tool layer. The AI is the product. That's the reverse of how every CRM before this has been built. YouSpot.com if you want to try it.

I found a way to see everyone who engages with my competitors' posts. here is the stack I used, in the order I ran it. I wanted a warmer list than a cold export. people commenting under posts in my category already care about the topic, and LinkedIn keeps those names public. so I wired six tools together and worked back from one post. - I start with phantombuster's activity extractor, pointed at a competitor's profile, to pull their recent posts with like and comment counts. that told me which post actually landed. it does not export who engaged, so that came next: https://lnkd.in/geMkAiDd - then the post commenters export. I run this one before the likers list, because someone who wrote a sentence is warmer than someone who tapped a button, and I get their actual words to open with. I kept it under 50 posts a day: https://lnkd.in/gkevdHkd - the post likers export widens it out. LinkedIn only displays 3,000 likers on a post, so that capped how far I could take any single one: https://lnkd.in/gtqZhBmb - clay's waterfall finds the work emails. it tries over 150 databases one after another until one comes back with a match, which stopped me losing rows when a single provider had nothing: https://lnkd.in/gC6eCRgU - claygent sits in that same clay table. I feed it the person's role and their comment, then ask it to score fit from 1 to 10 and draft one opening line quoting what they wrote: https://lnkd.in/giNDaUqE - instantly sends the sequence. every plan includes unlimited sending accounts, so I spread the volume and kept each domain under its daily limit: https://lnkd.in/gNRyyjbj I have run this against a handful of competitor posts now. the commenters list has out-converted the likers list every single time. happy to walk through how the clay table is wired if that is the part you want, just say so in the comments.

We found a way to see everyone who engages with a competitor's posts. Here is the stack we used, in the order we ran it. We wanted a warmer list than a cold export. People commenting under posts in a category already care about the topic, and LinkedIn keeps those names public. So we wired six tools together and worked back from one post. - We start with PhantomBuster's Activity Extractor, pointed at a competitor's profile, to pull their recent posts with like and comment counts. That showed us which post actually landed. It does not export who engaged, so that came next: https://lnkd.in/gGi637Nd - Then the Post Commenters Export. We run this one before the likers list, because someone who wrote a sentence is warmer than someone who tapped a button, and we get their actual words to open with. We kept it under 50 posts a day: https://lnkd.in/gcn-CNun - The Post Likers Export widens it out. LinkedIn only displays 3,000 likers on a post, so that capped how far we could take any single one: https://lnkd.in/gES9jRm8 - Clay's Waterfall finds the work emails. It tries over 150 databases one after another until one comes back with a match, which stopped us losing rows when a single provider had nothing: https://lnkd.in/gUNc8UEi - Claygent sits in that same Clay table. We feed it the person's role and their comment, then ask it to score fit from 1 to 10 and draft one opening line quoting what they wrote: https://lnkd.in/gfv-xs52 - Instantly sends the sequence. Every plan includes unlimited sending accounts, so we spread the volume and kept each domain under its daily limit: https://lnkd.in/g__x57VS At Supernodes we have run this against a handful of competitor posts now. The commenters list has out-converted the likers list every single time. Happy to walk through how the Clay table is wired if that is the part you want, just say so in the comments.

Last week I posted about how proud I was of my "wonderful" first Clay build. Multiple tabs, interdependencies, waterfalls, Claygent formulas, run conditions…all of it stacked on top of each other. BUT… Then I ran it against a small batch of low-priority contacts, and it was a MESS. I did what every first-time builder does: I overbuilt before I validated anything. Then, I tore it apart and rebuilt it by hand, with far less help from Claude the second time around. It was BRUTAL…but the 2nd time was when I really learned Clay. There are lessons buried in Techstars' book "Do More Faster" that I read years ago that I still think about: don't overbuild, start minimal and fail fast. Brother Tom Fahey at Saint John's High School used to say the same thing in four words: "Keep it simple, stupid (KISS)!" One step at a time... I learned a version of this in grad school too… A professor explaining business expansion put it this way: Change ONE dimension at a time. Either: A) New product OR B) Different geography. Never both at once, because you can't tell which change actually worked or broke. The dopamine hit from adding more: more enrichment, more sophistication, more scope, isn't the same as building something that actually works. Validate one thing. Keep it boring, one step at a time, and validate. Then earn the right to add the next one. Can you think of any other domains in life where it’s hard to restrain yourself from “overbuilding” before you’ve earned the right?

“CoLd eMaiL tO E-cOmMeRce is DeAd!” ok. here’s how we generated 45 leads in 20 days for this ecom agency: 1. Create a STRONG frontend offer (free asset/work) Everyone is spamming e-com with the same crappy pitches. “We guarantee to increase email revenue by 50% in 90 days or u don’t pay!” “We’ll create a landing page that increases conversions by 20%!” “We’ll 2X your ROAS or I’ll buy you a steak dinner!” This slop has been pushed since 2022, THIS outreach is dead. What’s not dead is actually offering something people want, regardless how saturated the industry is. What that looks like: - Free case study masked as a “playbook” - Free sample of your product/service - Free micro service that isn’t your main service (just a lead magnet) These work, and brands still want them when positioned right. 2. Hyper-qualify lead lists E-commerce data is typically pretty bad for some reason across most databases. Even when you filter for 11+ employees, you’ll run into brands that somehow have less than a $1k/mo budget for marketing spend lol. So you need to qualify further than the base filters: - Run claygent to confirm is an Ecom brand - Score the brands online presence (signal whether have budget or not) - Filter for E-com technologies (Shopify, WooCommerce, etc) - Verify whether they’re actively running Meta , Google, LinkedIn, Apple Ads or not with Apify Apollo.io alone is not going to give you a well targeted list, need to qualify further. 3. Very short & casual messaging Compared to most industries, E-commerce is probably the most “laid back” when it comes to messaging formalities. From what we’ve seen, the more formal you are the worse emails perform. As well as email length – these guys are receiving 20-100 cold emails/day, they need to be able to understand what your entire email is about in <5 seconds or they’ll just skip it. The best performing email we’ve ever sent to E-com brands was a one-line email that is less than 20 words. It’s along the lines of “John, interested in XYZ thing I created for your brand?” Print leads and meetings, which really came down to the offer. Message just presented the offer in a quick and simple way. — Email to ecommerce is really not the beast a lot of people in the space make it out to be, there’s just less margin for error. Follow these frameworks and you’ll be golden.

I counted the platforms we touch to get one GTM outcome. Five. Claude, re-prompted because the answer wasn't quite right. Cursor, writing into repos nobody has fully mapped. Zapier, holding a workflow together that one person understands. Claygent, enriching records nobody has audited. LangChain, running agents that don't talk to each other. Each one is good at its own job. The problem lives between them. None of them share memory, so context you established in one is invisible to the next and you rebuild it by hand every time. That rebuild is unpaid work nobody puts on an invoice. Add up the subscriptions. Then add the hours spent moving context between them. That second number is the real invoice.

At our recent NYC meetup with Clay, Head of AI Jeff Barg, ML Engineer Vyshnavi Khota and Software Engineer Soroush Khadem shared how they scaled agent evals to 300 million runs a month. Watch the full talk to learn: ✅ How Claygent and Sculptor got to production scale ✅ Clay's four quadrant framework for agentic evals ✅ Why closing the loop between production and offline evals is the hardest part ✅ How a data lake and long context are changing what agents can do with data Check it out on YouTube: https://lnkd.in/grStyZ4m

Before buying another AI sales platform, choose one bottleneck and one tool. - Slow account research: test Claygent. - Weak email drafts: test Lavender. - Missed reply handling: test Make with an OpenAI classifier. - Incomplete call notes: test Fathom. - Poor prioritization: test HubSpot Lead Scoring or Common Room. Run the experiment for 30 days. Measure the baseline, introduce the tool into one defined step and review a human sample. The metric must follow the work: valid accounts per hour, positive-reply rate, speed-to-lead, missed opportunities or completed next actions. More output is a productivity gain. Revenue operations needs the downstream metric to improve.

I built a fully automated B2B outbound engine inside Clay from scratch, taking raw firmographics all the way to verified decision-makers and custom cold emails. To make the scenario realistic, I picked an existing YC-backed developer tool as my benchmark case study. (Vendo (YC S26) it is) Here's how the end-to-end system works and why I structured the scoring logic this way: → Target Definition: Sourced US-based B2B SaaS accounts (11–200 employees) experiencing feature backlog fatigue and limited engineering bandwidth. → Contact Waterfall: Built a dynamic buyer hierarchy hunting for Product Leaders (Head/VP of Product) first, falling back to Engineering Leads, and defaulting to Founders, merging them into a single clean Decision-Maker column with verified LinkedIn URLs. → Custom Scoring Engine: Designed a multi-layered formula matrix where every weight directly reflects technical fit and intent: • Size Score (+2/1): Weighted 50–200 employees higher (+2) due to immediate request volume, while 11–49 (+1) signals high growth potential. • Industry Score (+2): Prioritized B2B SaaS & DevTools where in-app customizations are a core requirement. • Core TechStack (+2): Targeted modern frontend frameworks (React, Next.js, Vue, TS) where Vendo (YC S26)'s UI embeds seamlessly. • API & Infra (+1): Verified modern backend infrastructure (GraphQL, OpenAPI, Vercel) ready to power user-built micro-apps. • Integration Score (+2): Detected active marketplaces/directories using Claygent, a direct proxy for custom request friction. • Hiring Score (+2): Scraped roles for Product Managers, Solutions/Integration Engineers to identify teams spending heavily to manage client requests manually. • Funding Score (+1): Flagged recent capital raises to identify teams with budget and active growth mandates. → ICP Tiering & Validation: Summed all weighted signals to categorize accounts into Hot (>=10), Warm (7-9), and Cold (<7) buckets, verifying emails via Enrichley + MX domain checks. → Strategic Copy Generation: Generated 2-paragraph cold emails (under 90 words, zero sales fluff, no em dashes) paired with dynamic binary CTAs, driven by separate Claygent prompts for body and subject lines. → Dual-Table Architecture: Offloaded heavy scraping and formulas into a backend Lookup Table, pushing only 4 clean fields (Domain, ICP Fit, Subject Line, Email Body) to the primary execution view. Recorded a raw Loom walking through the entire table architecture step-by-step. Link is in the comments. I'm actively taking on GTM Engineering projects. If you're looking to automate your outbound stack from the ground up, let's connect via DM or 📩 workwithdeblina@gmail.com Nour Zahzah Yousef Helal, would love to know what you think of this setup for Vendo (YC S26)! And a special shoutout to Yogesh Jaiswal for always guiding me and having my back.

Two archetypes, daily workflows, real tools, salary ranges, and when to hire a GTM engineer vs automate the role with Deepline + Claude Code.

میں نے سجدے میں روتے ہوئے کہا، اللہ پاک کیا آپ مجھ سے ناراض ہیں؟ میرے اللہ نے جواب دیا: میں جن سے ناراض ہوتا ہوں، انھیں یہ فکر کہاں ہوتی ہے۔ ❤️🥺" ​#IslamicQuotes #Namaz #Allah #DeepLines #HeartTouching #IslamicShorts #Sukoon #QuranicQuotes #StudiousVibes ​اسلامی کوٹس، نماز کی باتیں، اللہ کی محبت، دل کو چھو لینے والی باتیں، اسلامی ویڈیوز، سجدے کی حالت، سکونِ قلب، اسلامی اسٹیٹس، Urdu Quotes, Allah Loves You

What happens after the outreach starts? GTM Engineering is not just about finding leads. It’s about connecting the entire system. As I have been learning GTM Engineering, Ihave worked on company discovery, enrichment, buying signals, scoring, Clay workflows, APIs, MCPs, and AI-assisted outbound. But one question kept becoming more important: What happens after the outreach starts? That’s where OutboundSync caught my attention. OutboundSync connects sales engagement platforms like Instantly.ai, Smartlead, HeyReach and EmailBison with CRMs like HubSpot, Salesforce, Close and Attio. It can sync outbound activity, replies, meetings and social activity into the CRM, while also supporting webhooks, APIs, filtering, routing and AI agent skills. For me, this adds another layer to the GTM engineering workflow: Discover → Enrich → Score → Outreach → Sync → Analyze → Optimize It means I can build systems where outbound data doesn't get trapped inside separate tools, but becomes useful throughout the revenue workflow. The Bootprint 🥾 (prev Clay Bootcamp) Scholarship will give me the opportunity to keep developing these skills and become a stronger GTM Engineer. 🙏 Thank you to OutboundSync for sponsoring the Bootprint 🥾 (prev Clay Bootcamp) Scholarship and supporting the next generation of GTM Engineers. Big thanks also to EmailBison Girls Who Clay Deepline Clay Nathan Lippi 🥾 FastForward RevenueHoop What part of your GTM stack would you most like to automate or connect?

I get my leads with Deepline. Come check it out at our Run Hack this Saturday: https://luma.com/tag-wx1t Btw, all runners have access to Deepline during the Run Hack. Lucky them ;) Shoutout to my fave actors: Fergus McKenzie-Wilson Tijs Nieuwboer Luke Balabanovic Zakee Abdi

Head-to-head comparison of Clay, Apollo, and Deepline for B2B enrichment workflows. Real pricing, data coverage, and agent support compared.

112 opportunities created in 39 days. For an anonymized PE client: → 11.3K targeted emails → 7.29% reply rate → 112 opportunity paths created The actual stack behind the work: AI Ark + LimaData for the data foundation. Deepline, Firecrawl, Exa, Codex, and Claude to turn company signals into relevant context. n8n to route qualified replies into the right next action. The message was never the strategy. It was the final mile of a thesis-led origination system.

How can you run and build at the same time ? One Answer : Cognition x The Interaction Company of California We got you the best tools to both build and run at the same time for the First Running Hackathon in Europe So what you waiting for to apply ;) 🏄🏼‍♂️ : https://luma.com/tag-wx1t We got you Wispr Flow Cognition x The Interaction Company of California , ElevenLabs , Healf , ROXFIT , algosoup , Delfa , Deepline , Thrad , Unicorn Mafia , Security Builders Club, Accelerate ME, O2 (Telefónica UK), Tavily, Pitchless Community, PerfectTed

Another successful utility locating project completed in Delray Beach, Florida! 📍 Using Ground Penetrating Radar (GPR), DeepLine FL accurately located and marked underground utilities, helping our client work safely and avoid costly damage. We find the path before you drill. 📞 (754) 272-7401 ✉️ sales@deeplinefl.com 🌐 www.deeplinefl.com #DeepLineFL #GPR #UtilityLocating #GroundPenetratingRadar #DigSafe #SouthFlorida

For junior / new GTM engineers, the market is quite interesting right now, but im seeing a problem: There has never been more demand out there, but you can no longer sell "cold email" or "outreach" as as service, nor "revops" and AI has made operations much easier & tools are cheaper than ever. So: You have to sell the outcome; and to sell the outcome though you have to understand your buyer. They're not "struggling with pipeline", they just build random lists, because they don't have their TAM mapped out. They're not "struggling with closing", they just have no system in place to derive meaning & learn from their outbound & sales calls, so their understanding of their buyer does not improve over time and the Clay hype is dying down; a GTME is not new anymore as an idea. Most companies still don't know how to scope a GTM role, but at least they know it exists. So they want it IN HOUSE. That means that while demand is going 📈 , companies are pushing hard on in house only. This means that you can either go in house, which means risking becoming too one sided or falling behind if the company is not AI native or staying in the Agency / Fractional GTM world, which means you have to become great, since good won't cut it anymore. Companies like Deepline & prospeo.io have largely democratized the infrastracture/data part of GTM & the existence of MCPs means that the "executor" GTM people / juniors are essentially useless right now, since a properly scoped task by a senior GTM person, who knows how to do strategy, needs no executor, since an agent can do all the execution & implementation. So that leaves us with 2 important trends for GTM engineers: - You arguably have to aggressively upskill to become senior enough not to get crushed by agents in 6-12 months. - You have to either go inhouse and accept the risks & benefits OR stay fractional/agency and accept that only the top half will have a chance to thrive Greetings from Athens, GR 👋

A practical workflow for using Deepline skills to validate technographic signals before deploying scripts and Workflows.

Most people use AI at the surface and miss the system underneath. 𝗕𝗲𝗰𝗼𝗺𝗲 𝗮 𝗰𝗿𝗲𝗱𝗶𝗯𝗹𝗲 𝗖𝗵𝗶𝗲𝗳 𝗔𝗜 𝗟𝗲𝗮𝗱𝗲𝗿. 𝗘𝗮𝗿𝗻 𝘁𝗵𝗲 𝗔𝗜 𝗰𝗲𝗿𝘁𝗶𝗳𝗶𝗰𝗮𝘁𝗶𝗼𝗻𝘀 𝗯𝘂𝗶𝗹𝘁 𝗳𝗼𝗿 𝗹𝗲𝗮𝗱𝗲𝗿𝘀. 𝗚𝗲𝘁 𝗔𝗜 𝗰𝗲𝗿𝘁𝗶𝗳𝗶𝗲𝗱 𝗻𝗼𝘄 → chiefaileader.com/ip __________ Here is the simplest way to understand what AI is built on. Think of AI like an iceberg. Above the water: → ChatGPT → Claude → Midjourney → Gemini That is what everyone sees. But the useful part is underneath. ✅ Layer 1: Classical AI This is the 1950s foundation. Symbolic AI, expert systems, knowledge representation, logic, and reasoning. Diagnostic check: if a workflow is rules-based, it may still live here. ✅ Layer 2: Machine Learning This is where systems learn from data instead of fixed instructions. Supervised learning, unsupervised learning, classification, regression, reinforcement learning. Consequence: better data often beats a better prompt. ✅ Layer 3: Neural Networks This is where connected nodes learn patterns. Perceptrons, cost functions, activation functions, hidden layers, backpropagation. Action: learn backpropagation once and modern AI becomes less mysterious. ✅ Layer 4: Deep Learning This is where scale changed the game. Transformers, CNNs, RNNs, LSTMs, autoencoders. Before: narrow pattern recognition. After: language, speech, images, and code became practical. ✅ Layer 5: Generative AI This is the layer most leaders now recognize. LLMs like GPT, Claude, and Gemini. Diffusion models like Midjourney and DALL-E. Multimodal models that combine text, image, audio, and video. ✅ Layer 6: Agentic AI This is the 2026 shift. Memory, planning, tool use, and autonomous execution. AI does not just answer. It starts acting across workflows. One simple leadership test: → Ask which layer each AI initiative depends on. → Ask what data, tools, and approvals it needs. → Ask where a human must stay accountable. If nobody can answer those 3 questions, the project is probably tool-led instead of strategy-led. The mistake is treating all AI as one thing. It is not. Different layers create different risks, skills, and opportunities. If you lead a team, this matters. You cannot build a strategy around a tool name. You need to know the layer underneath it. That is how you separate hype from useful adoption. Which layer does your team understand least right now?

Super interesting! "Artificial Intelligence and the Brave New World in Finance" by Markus K. Brunnermeier. "Economics assumes that even when people know different things, they have a common understanding and hence can describe the world to one another in a shared language. Trust in the understanding of experts, extended by institutions, gives each person access to a broader societal understanding. Society hence understands more than any of its members. The introduction of agentic AI is qualitatively different from previous innovations. It disrupts the trust arrangement, undermines societal understanding and leads to asymmetric understanding: AI agents can learn how humans think and respond, while humans may be unable to understand or reliably anticipate how those agents will act. This asymmetric understanding can make prices harder to read (less informationally efficient) and put central banks at a strategic disadvantage when engaging and communicating with market participants. Preparing for the asymmetric understanding scenario calls for segmented markets that preserve a human fallback, simpler and more robust central bank rules, and less strategic ambiguity" "For helpful comments and discussions, I thank Claude Fable from Anthropic..."😉 https://lnkd.in/ebJHQiwc

Some suggestions for when campuses should "build or buy" edtech software, based on my observations aross Idaho higher ed. This perspective comes from working within my own college but also what I learned from my work last year as the Idaho State Board lead for the AI Catalyst Initiative. This piece also came from my recent experience using Codex to develop my own mini agent-native textbook platform. It took a lot of time and tokens, definitely not free, but the benefit to me is enormous because of how it works with Codex and Claude Code. And yet, I think it's a terrible idea for institutions to develop their own publishing platforms! There are benefits to shared platforms. So when does it make sense to build or buy? It depends. What I see right now is that the most successful experiments with building rather than buying come from talented individuals who figure out how to solve a technical problem that allows their institution to become less reliant on enterprise partnerships. We all have access to Gemini or Copilot, for example, and yet there's very little we can do with our official tiers because of complicated permissions, authentication, etc. It's a nightmare trying to make some of this work. So Boise State University, University of Idaho, and other campuses have been developing their own AI tech stack to supplement. And it's entirely due to talented faculty & staff who have a knack for working with agents. Based on these examples and others I've observed, here are my rules for how and when to (vibe) build: 1. Invest in faculty and staff who have a knack for working with agents. 2. Go hard on vibe-coding in low-risk situations. 3. The best solutions often happen through faculty or staff prototypes that eventually get audited and perfected by the campuses software dev team. When not to build (or when to approach it delicately): 1. Software that involves PII & institutional data is a separate category of vibe-coding. 2. Some legacy and enterprise software is better just because it’s shared. This applies to Google docs as well as publishing platforms. There are clear benefits to shared infrastructure that's used widely by other institutions (and K-12). There's risk in becoming too bespoke. I also point out that you’re likely to see more “build it” experiments coming out of Idaho over the next few years in part due to Liza Long, Ed.D.'s efforts at the Idaho Office of the State Board of Education. In 2025, OSBE won a $4 million FIPSE grant, thanks to Liza’s proposal, and most of that is being passed along to the eight public higher eds across Idaho to fund AI-readiness efforts. Some of the best sub-grant proposals we’re seeing are strong examples of how our Idaho institutions are choosing to build local AI solutions. Jonathan Lashley Kevin Rank, MBA Phil Merrell

AI just got hands. Anthropic's new Model Hardware Standard could change how physical operations work forever. Anthropic built MHS to let AI agents control real physical devices like robotic arms, lab equipment, and multi-machine systems through a single unified interface. Early tests cut integration time from weeks down to hours. At QuEra, Claude ran 700 automated tests and succeeded 99.3% of the time with no language model in the loop. This is the same standardization logic behind their Model Context Protocol for software, now extended to the physical world. AWS, Doosan Robotics, Universal Robots, Hugging Face, and Raspberry Pi are already building MHS support. The catch? Claude still needs people in the loop. In one Genentech test, it kept restarting a failed process without recognizing the problem was physical, not software. Expert oversight is essential right now. Do this: Map one workflow this week where your team manually coordinates between two or more disconnected systems. That gap is exactly where AI agent protocols like MHS and MCP are headed next. What would you automate first if AI could coordinate all your disconnected tools and systems through a single interface? See how Leads to Conversion can help your business grow with AI -&gt; https://lnkd.in/etMZ5YVp https://lnkd.in/eA_znieP

Retractions have become common in academic research these days. In my opinion almost all authors will have more or less papers retracted in the next 5-10 years. ‎ ‎So the academic excellence will shift from having a retracted paper or not to the reasons and numbers of retractions. Interviews will assess honesty by acknowledging retractions rather than hiding it. ‎ ‎These retractions can be due to methodological errors, guideline violations, funding issues, ghost authorship, or fabricated/AI-generated data. ‎ ‎One important factor is fabricated information using ai tools like ChatGPT, Gemini, Claude etc. All as a result of the race to win more publications in less time with less efforts. A whole study built on fake patient data being utilized for real world patients. ‎ ‎Lets see where this trend leads to. What's your opinion on it? ‎ ‎From the platform of Klinital Healthcare

Tu empresa no necesita otro chatbot. Necesita un agente que haga cosas. Muchas empresas utilizan WhatsApp para atender clientes. El problema es que detrás de cada mensaje suele haber una persona que tiene que: → Leerlo → Entender qué necesita el cliente → Buscar información → Responder → Registrar los datos → Actualizar el CRM → Gestionar una cita → Hacer seguimiento Y repetirlo cientos de veces. Pero, ¿qué pasaría si WhatsApp pudiera convertirse en el punto de entrada de un proceso automatizado? MENSAJE ↓ WHATSAPP BUSINESS API ↓ AUTOMATIZACIÓN ↓ AGENTE IA ↓ CLAUDE ↓ DECISIÓN ↓ ACCIÓN El agente puede analizar la conversación y, según el caso: 🔹 Clasificar el lead 🔹 Recoger información 🔹 Resolver preguntas frecuentes 🔹 Consultar datos 🔹 Crear registros en el CRM 🔹 Gestionar reservas 🔹 Programar seguimientos 🔹 Avisar a un comercial 🔹 Derivar la conversación a una persona La clave está aquí: No queremos que la IA simplemente responda. Queremos que la IA entienda qué está ocurriendo y active el proceso correspondiente. Eso cambia completamente el concepto de automatización. WhatsApp deja de ser simplemente un canal de comunicación. Se convierte en una puerta de entrada a los procesos de la empresa. Y esto puede aplicarse a clínicas, inmobiliarias, talleres, empresas de servicios, asesorías, laboratorios, ecommerce y prácticamente cualquier negocio que gestione un volumen importante de conversaciones. En Beltranintelligence no buscamos poner IA porque sí. Primero detectamos qué proceso está consumiendo tiempo. Después diseñamos la automatización. Y finalmente conectamos la IA con las herramientas que ya utiliza la empresa. Menos tareas repetitivas. Menos errores. Más capacidad operativa. Si tu empresa recibe muchos mensajes por WhatsApp y todavía se gestionan manualmente, probablemente exista una oportunidad de automatización delante de ti. 💬 Si quieres, escríbeme “WHATSAPP” por privado y te explico qué procesos podríamos automatizar en tu negocio. #InteligenciaArtificial #Automatización #AgentesIA #WhatsAppBusiness #IAEmpresarial #N8N #TransformacionDigital #PYMES #AutomatizacionEmpresarial

The UI collapsed. The workforce is next. Two years ago at the AWS summit, I said that complex enterprise UI would collapse into a single text box. Any operation, any query, one conversation. It was not a popular claim. On August 26, Salesforce and Anthropic announced #Claudeforce. The largest CRM in the world put its data, workflows and governance inside #Claude. A seller can now prep a meeting, review deal health and update pipeline without ever opening Salesforce. Marc Benioff summarized it in one line: the UI is the AI. So here is my next claim, and it is the bigger one. Businesses will no longer be powered by humans. They will be powered by an autonomous workforce, with humans setting the standard and holding the judgment. Look at what #Claudeforce actually shipped. A person asks. The AI reasons and acts. The person approves. The human is still in front of the work, so the organization can still only produce as much as its people have attention to drive. That is a faster tool, not an autonomous workforce. The difference is the whole economics. A tool multiplies a person while the person is working. An autonomous workforce does the work whether the person is in the room or not. Almost everything the market calls AI transformation sits on the first side of that line, which is why more than 80% of companies report no bottom-line impact from AI while adoption keeps climbing. The next move is not more capability at the interface. It is the human stepping behind the work: teaching the activity, delegating it, then supervising. What gets handed over is not a prompt. It is responsibility for the work. #Claudeforce proves the model is no longer the constraint. Frontier reasoning can now act inside enterprise systems under enforced business rules. The constraint is who can teach the work. Salesforce and Anthropic authored those 37 sales skills themselves. A strong answer for sales from two of the best engineering organizations in the industry. No answer at all for the operations manager who owns a returns exception process, or the support lead whose expertise is the twelve edge cases that never made it into an SOP. They cannot wait for someone to author a skill for them. They are the only ones who know what it should contain. My claim is that the autonomous business belongs to the people who know the work. On September 10 I am speaking at the Amazon Web Services (AWS) Summit in Tel Aviv with Yariv Nir the CIO of Tel Aviv Sourasky Medical Center, who is building an Agentic Hospital OS. Clinical and administrative work that runs on its own, in one of the busiest hospitals in Israel. At AllCloud we are helping customers build these autonomous capabilities using Amazon Web Services (AWS), Salesforce and Anthropic. Eran Gil, Ronit Rubin, Raz Dar, Amir Hunga, Felix Marnin, Itzik Amar, Omri Kovalski, Jonathon Nisenboum, Peter Nebel, Gabriel Romero, Hans Schabert, Ilan Froimovici, Kiran Randhi, Uri Tamir

I built a Claude agent that runs your entire LinkedIn content system. I put together the set up on a document and I'm giving it away 👇 You give it your LinkedIn profile once. It handles the rest every week. I packaged the whole pattern. 7 tools, free: → The voice profiler, so your posts sound like you and not a template → The daily trend scanner pulling from Reddit and X, so you're never writing into a dead topic → The content matcher that maps your ideas to 12 proven post formats, with image direction included → The post writer that drafts full posts, lead magnets and case studies in one pass → The performance tracker that logs what books calls and adapts any post in one click, so you stop guessing what to write next → The lead magnet builder that creates the actual resource, not just the post about it → The scheduling engine that tells you exactly when to post each piece 5 minutes to set up. 15 minutes a week to run. I'm giving free access to my setup.

🚨 أهم تحديثات الـDigital Marketing هذا الأسبوع تحديثات هذا الأسبوع من Meta وInstagram وTikTok وGoogle وLinkedIn وYouTube وOpenAI، والاتجاه واضح: AI + Automation + Commerce + Content Quality ⬅️ Instagram يسرّع صناعة الـReels ميزة First Draft تساعد على إنشاء مسودة أولية للـReel من فيديوهاتك، مع اختيار وقص الأجزاء المناسبة. Instagram أكدت أيضاً أهمية الـCaptions للـSEO: Hook + Keywords + CTA. ⬅️ Meta تدفع نحو الـAutomation بدأت Meta في تقليل بعض خيارات التحكم اليدوي في الـPlacements، والاعتماد أكثر على الـAI في اختيار أماكن عرض الإعلانات. التركيز أصبح أكثر على Creative + Offer + Data. ⬅️ Auto-DM لإعلانات Instagram Meta تختبر إرسال DM تلقائي لمن يكتب Keywords معينة في تعليقات الإعلانات. فرصة قوية للـLead Generation وConversion. ⬅️ TikTok يدخل الـDigital Out-of-Home توسّع TikTok شراكاته مع الشاشات الرقمية في المطارات والشوارع ومحطات الوقود والمتاجر، مع إدارة الحملات من Ads Manager. TikTok لم يعد داخل شاشة الهاتف فقط. ⬅️ TikTok يقترب من In-App Commerce اختبار توسيع TikTok Pay للدفع والتحويلات أثناء الـLive. Content → Product → Checkout بشكل أقصر = Friction أقل وConversion أكبر. ⬅️ LinkedIn يحارب AI Slop أكثر من مليون بلاغ ضد محتوى AI منخفض الجودة، مع تراجع وصوله. المشكلة ليست AI، بل محتوى بلا Value أو Experience أو Opinion. ⬅️ Google Ads تسهّل الـA/B Testing تجربة أبسط لمقارنة أداء الحملات التقليدية مع AI Max قبل توزيع الـBudget بشكل أكبر. ⬅️ Claude أصبح يتذكر سياقك Anthropic أضافت Memory موسعة، ما يسمح بالاحتفاظ بسياقات سابقة. للـMarketers: يمكنك بناء Context حول Brand Voice + Audience + Strategy واستخدامه لاحقاً. ⬅️ X يفتح الباب أمام AI Agents إطلاق MCP Server لربط أدوات الـAI بالحسابات وأتمتة تحليل المحتوى والـTrends والأداء. الاتجاه ينتقل من: AI يجيب → AI يفهم ويحلل وينفذ. ⬅️ YouTube وThreads يتوسعان في الـPodcasts YouTube يختبر Stations لتجربة Audio/Music مستمرة، وThreads يختبر Transcripts متزامنة مع الصوت. فرصة لإعادة توزيع الـAudio على Formats متعددة. ⬅️ WhatsApp يرفع مستوى الأمان تحديثات تشمل Passkey وتحسين Two-Step Verification. مهم للـBusinesses التي تعتمد على WhatsApp في Sales. ⬅️ Fake Engagement قد يضر حسابك Meta حذرت من شراء الـLikes والـEngagement. Vanity Metrics ≠ Business Results الـ10,000 Like المزيفة لن تعطيك بالضرورة Leads أو Sales. ⬅️ OpenAI تتجه نحو AI Agents التحول الأكبر: من AI كأداة إلى AI قادر على تنفيذ مهام كاملة مثل تحليل البيانات وتنفيذ Workflows. 🎯 الخلاصة: الـDigital Marketing يتحرك بسرعة نحو: AI + Automation + Commerce + Content Quality والميزة لن تكون لمن يستخدم AI أكثر… بل لمن يعرف كيف يحوّله إلى نتائج حقيقية. ↩️ تابعنا دائماً لأهم تحديثات الـDigital Marketing أولاً بأول.

Retractions have become common in academic research these days. In my opinion almost all authors will have more or less papers retracted in the next 5-10 years. ‎ ‎So the academic excellence will shift from having a retracted paper or not to the reasons and numbers of retractions. Interviews will assess honesty by acknowledging retractions rather than hiding it. ‎ ‎These retractions can be due to methodological errors, guideline violations, funding issues, ghost authorship, or fabricated/AI-generated data. ‎ ‎One important factor is fabricated information using ai tools like ChatGPT, Gemini, Claude etc. All as a result of the race to win more publications in less time with less efforts. A whole study built on fake patient data being utilized for real world patients. ‎ ‎Lets see where this trend leads to. ‎ ‎From the platform of Klinital Healthcare.

Build Your Claude Workflow in Just 5 Minutes 𝗕𝗲𝗰𝗼𝗺𝗲 𝗮 𝗰𝗿𝗲𝗱𝗶𝗯𝗹𝗲 𝗖𝗵𝗶𝗲𝗳 𝗔𝗜 𝗟𝗲𝗮𝗱𝗲𝗿. 𝗘𝗮𝗿𝗻 𝘁𝗵𝗲 𝗔𝗜 𝗰𝗲𝗿𝘁𝗶𝗳𝗶𝗰𝗮𝘁𝗶𝗼𝗻𝘀 𝗯𝘂𝗶𝗹𝘁 𝗳𝗼𝗿 𝗹𝗲𝗮𝗱𝗲𝗿𝘀. 𝗚𝗲𝘁 𝗔𝗜 𝗰𝗲𝗿𝘁𝗶𝗳𝗶𝗲𝗱 𝗻𝗼𝘄 → chiefaileader.com/bp Original post: __________ Most people don't get bad results from Claude because AI is limited. They get bad results because their prompts are unclear. A better prompt creates a better outcome. Here's how to master Claude in 3 minutes: Step 1. Start with a clear prompt structure. Give Claude 5 key details: → Task: What do you need? → Context: What background should it know? → Format: How should the answer look? → Tone: How should it sound? → Goal: What result do you want? Step 2. Be specific. "Write about marketing" is too broad. Instead: "Explain 3 simple marketing strategies for small businesses using basic English with examples." Clear instructions lead to stronger outputs. Step 3. Break big tasks into smaller steps. Don't ask Claude to do everything at once. Start with: → Generate ideas → Improve the draft → Adjust the tone → Create the final version Better process = better results. Step 4. Define your output format. Tell Claude exactly what you need: → Bullet points → Tables → Short paragraphs → Step-by-step guides → Scripts Don't make AI guess your preferred structure. Step 5. Give examples. Want a specific style? Show Claude what good looks like. Examples help it understand your: → Voice → Structure → Writing style Step 6. Control the tone. Your first response is only a draft. Ask Claude to: → Make it shorter → Simplify the wording → Add examples → Make it more engaging Step 7. Use the "Improve this" method. Paste your content and ask: "Improve this. Make it clearer, more engaging, and easier to read." Small improvements can create better results. Why people struggle with AI: They treat prompts like quick commands. They don't provide enough context. They expect perfect answers without refining. AI works better when you treat it like a collaborator. The quality of your input shapes the quality of your output. Credit to kumkum Shinde. Follow her for more.

What if an AI agent could make sure no promising sales lead ever goes cold? 👀 Day 1 of the Claude Hackathon ’26 by Product Space! 🤖 Today, I started exploring a problem that sounds simple but can cost sales teams real opportunities: Sales reps have conversations all day — but important follow-ups can easily get forgotten. A prospect asks for pricing. Someone promises to send a case study. A follow-up is agreed for next Tuesday. Then dozens of other conversations happen… and the opportunity gets missed. 💡 The idea I'm building FollowUpAI — an AI sales follow-up agent that can: 🔹 Analyze sales calls, emails, and meeting notes 🔹 Identify prospects who need attention 🔹 Detect promises and missed follow-ups 🔹 Prioritize leads based on urgency and buying intent 🔹 Recommend the next best action 🔹 Draft personalized follow-up messages 🔹 Maintain a prioritized list of opportunities The goal isn't to build another chatbot. I want the agent to understand → decide → recommend → act. 🧠 Today's biggest insight The interesting part of an AI agent isn't simply getting an answer from AI. It's giving AI enough context to answer: "What should happen next?" That's the direction I'm taking with FollowUpAI. Tomorrow: turning the idea into a working agent and getting the first end-to-end workflow running. What sales task do you think an AI agent should handle automatically? I'd love to hear your ideas 👇 #ClaudeHackathon #AI #AIAgents #BuildInPublic #Claude #SaaS #SalesAutomation

**Your next career move could be one share away.** We’re hiring across technology, data, AI, cloud, enterprise applications, finance, and support. If you’re exploring a new opportunity—or know someone who is—please share this post and connect with us. **CHENNAI / GDC CHENNAI** • **241596** — .NET + Azure + Angular | 10–15 years | Up to ₹25 LPA• **242729** — Salesforce Technical Lead• **243106** — Salesforce Marketing Cloud• **243140** — ELK Developer• **243552** — SAP S/4HANA EWM Consultant• **243941** — SAP ABAP CDS• **244146** — Java, Spring Boot, React, PostgreSQL, GCP• **244508** — Workday Extend & Orchestration Developer• **244509** — Senior Workday Integration Developer / Consultant• **244559** — Senior Workday Core HR Developer• **244594** — AWS Connect + Claude AI Developer• **244595** — ETL QA Analyst• **244592** — Endpoint + SCCM + JAMF + Intune **HYDERABAD** • **242505** — Product Owner• **242573** — Guidewire Technical Architect — Policy & Billing | Hyderabad / Bengaluru• **243548** — Oracle Fusion Technical Architect — OIC, BIP & VBCS• **243547** — Oracle Cloud OM Functional Architect• **244459** — AI/ML IoT Engineer — SQA• **244623** — Senior ETL Developer• **244627** — Business Analyst — Life Insurance **BENGALURU / BANGALORE** • **243833** — ServiceNow Architect• **244063** — SME — SAP Tax• **244066** — SME — SAP VIM• **244069** — SME — SAP Vistex• **244071** — SME — SAP BTP Admin• **244527** — C1 Analyst — IT Cloud COE• **244676** — IBM Planning Analytics Developer **MUMBAI** • **244271** — AUM — Associate Director Level• **244274** — Junior GC Accounting• **244277** — Fixed Assets & Billing Accounting — Associate• **244330** — Oracle Fusion Cloud EPM/EPCM Functional Consultant — Cost Allocation **PUNE** • **244276** — Accountant — Mid Level• **244275** — Senior Accountant — International **ABU DHABI** • **242512** — AI Engineer• **244424** — SAP VIM Consultant• **244693 / 244694** — Forward Deployed Engineer• **244695** — AI Value Architect **REMOTE** • **244293** — Integration Developer — CP4I **DUBAI** • **244619** — Full-Stack React Developer with .NET Backend — Airlines | 5+ years• **244621** — OPS — Cargo Application Support **Multi-location openings** • **244393** — Data Engineer — AWS | Chennai / Hyderabad• **244726** — Databricks ETL Engineer | Bengaluru / Bangalore Interested candidates can email their updated CV to [**chetan@kiashsolutions.com**](mailto:chetan@kiashsolutions.com). Please mention the **Job ID and designation** in the subject line. Know the right person? Tag them, share this post, or send it directly. One share can connect the right candidate with the right opportunity. #Hiring #TechJobs #ITJobs #Recruitment #CareerOpportunities #CloudJobs #DataEngineering #AIJobs #SAP #Oracle #Salesforce #ServiceNow #Workday #BengaluruJobs #ChennaiJobs #HyderabadJobs #MumbaiJobs #PuneJobs #AbuDhabiJobs #DubaiJobs #permroles #fulltimejobs #bengaluruopenings #chennaiopenings #hyderabadopportunities

For a small local business, these are good results: 100% ROI with $160,000+ revenue generated. Leads from AI platforms like ChatGPT, Claude, and Gemini have been the highest paying customers for my dental client. Every single lead from AI that has contacted the business has converted, and what's most interesting is that they fully trust the results and recommendations coming from the AI platform. Of course, the reputation and experience of the dentist are essential to the recipe.If you have 100+ 5-star reviews on Google, decades of experience, an optimized website, and more, you have a real chance of getting cited by AI. My dental client is a small local practice in Southern California. In the past 12 months, they've pulled in $160,000+ in revenue strictly from a handful of AI conversions with 100% ROI. This is a small cosmetic dental practice, and the leads are local. This is evidence that AI doesn't reward business size or ad spend. AI search rewards credibility. The key things I mentioned above all come back to EEAT: Expertise Experience Authority Trustworthiness There's an overwhelming influx of AI trends and "techniques" right now, but foundational SEO should still be the focal point of your digital marketing, regardless of how big or small your AI tech stack is. You can check for traffic from AI platforms, for free, in GA4. Go to Reports -> Life cycle -> Acquisition -> Traffic acquisition to get an idea of where you're at. #GEO #AEO #SEO #DigitalMarketing #DentalMarketing #EEAT #AISearch

watched "from ai-assisted to ai-native" by Clare Liguori, aws she leads kiro. her talk: how amazon builds now 6 engineers. 76 days. mantle: new inference engine for bedrock manual code: 1-2% of the codebase the rest: autonomous agents specs -> orchestrate -> review real artifact isn't code it's the harness. factory that produces code: loops, tools, steering, mocks, evals 2 distinguished engineers 1 senior principal 3 principals amazon's top tier > trap everyone sees "6 engineers" and copies the shape those six can build the factory alone they absorb qa, security, eval, platform regular teams can't absorb all these roles so they rely on tests tests pass. but is the code correct? tests: unit, integration (does this code work?) evals: golden tasks, quality metrics (does our harness work?) reviewers: PR review, architecture, business fit (are we building the right thing?) regular teams need explicit roles: owner. qa. eval. security. platform. harness is now a single point of failure with these roles it's owned and checked without them it's an orphan > tooling owning the process isn't enough if the runtime is rented claude code, codex, kiro you configure. you don't reshape you need a hackable harness: pi — everything is an extension deepseek harness — everything is a plugin your loop. your tools. your evals. your steering. if you can't hack the harness, you don't own the factory ai is a multiplier it multiplies the foundation no foundation = it multiplies your errors, at scale what's your foundation?

KI in Österreich: Es geht nicht mehr nur um Chatbots. Mit ida für Verwaltungsfragen und GovGPT innerhalb der Bundesverwaltung wird gerade sehr gut sichtbar, wohin sich KI entwickelt: Nicht noch ein weiteres Tool zum Ausprobieren – sondern KI als Bestandteil konkreter Prozesse. Genau das ist meiner Meinung nach auch für Unternehmen entscheidend. Viele Firmen beginnen bei der falschen Frage: „Welche KI sollen wir verwenden – ChatGPT, Claude oder Gemini?“ Die bessere Frage lautet: „Welchen Prozess möchten wir verbessern?“ Zum Beispiel: – Kundenanfragen vorsortieren – E-Mails und Dokumente vorbereiten – internes Wissen schneller zugänglich machen – Leads strukturieren – CRM-Prozesse automatisieren – wiederkehrende Aufgaben zwischen Website, E-Mail und internen Systemen verbinden Erst danach sollte entschieden werden, welches KI-Modell oder welche Plattform dafür geeignet ist. Auch international sieht man gerade dieselbe Entwicklung: ChatGPT wird stärker zur Arbeitsplattform, Gemini verbindet sich mit immer mehr Diensten und KI-Agenten bekommen zunehmend Zugriff auf reale Tools und Prozesse. Der eigentliche Fortschritt liegt deshalb nicht nur in besseren Modellen – sondern in besseren Integrationen. Für Unternehmen in Österreich bedeutet das: Man muss nicht gleich ein riesiges KI-Projekt starten. Oft reicht ein klar definierter Prozess, bei dem heute unnötig Zeit verloren geht. Ich habe die aktuellen Entwicklungen in Österreich und international in einem neuen KI-News-Überblick zusammengefasst: KI-News August 2026: Was sich in Österreich und weltweit verändert → https://lnkd.in/dbyy2tX2 #KünstlicheIntelligenz #KI #Automatisierung #Digitalisierung #Österreich #KMU

There is a limited window to become part of the stack the next generation of AI-driven companies builds on. Colin Nederkoorn’s post this week is one a lot of SaaS teams need to deeply consider. I'm not talking about adding another “AI assistant” inside your product. It is about becoming the thing an agent can set up and operate without friction. That is not a panacea. Distribution, trust, reliability, and price still matter. But the buying motion is changing faster than most roadmaps are keeping up with. A few years ago, a founder or lead compared tools, sat through a demo, and configured the integration. Now a lot of net-new stack decisions happen inside an agent loop. The agent looks for a clean API, an official MCP server, skills for coding agents, or a connector in ChatGPT or Claude Cowork. Then it moves on. The decision is made. That is why certain products keep showing up as defaults: 🎯 Resend made email something an agent can send, read, and administer 🎯 Vercel made the serverless hosting and agent / AI SDK layer the path of least resistance for shipping 🎯 Neon made Postgres something an agent can provision, branch, and query 🎯 PostHog made analytics, flags, and debugging available from the editor 🎯 Customer.io is making the same bet on the messaging layer This also extends past engineering teams. Knowledge workers are doing more of their work inside ChatGPT and Claude Cowork. If your product cannot be reached from those surfaces, you are not in the new default. You are an island, and there is a lot of friction to get there. What are some other examples of products that are showing up as defaults with AI driven buyers and implementors?

🚨 OpenAI just cut off Cursor. But the bigger story isn’t Cursor. It’s that AI model access is becoming strategic infrastructure. OpenAI says it will wind down its model-supply agreement with Cursor after SpaceX acquired Anysphere, with the current cutoff date set for November 12, 2026. Why does that matter? Because this is no longer simply: App → API → Model The AI stack is becoming: Models + Compute + Applications + Developer Workflows + Data + Distribution And when ownership changes, the relationship between those layers can change too. 🔄 Here’s the part I find most interesting: Cursor says OpenAI models account for only about 5% of its traffic. So this isn’t really a 5% story. It’s a control story. 🔐 OpenAI is willing to give up direct supply rather than keep its frontier models inside a platform now controlled by a major rival. Anthropic, meanwhile, is scaling capacity for Claude in Cursor. And developers still have alternatives such as BYOK and Codex. That leads to a bigger question for every enterprise building with AI: What happens when your “best model” is no longer available through your preferred platform? Because acquisitions, contract changes and competitive conflicts can suddenly become architecture problems. The answer may be a shift from: ❌ Single-provider dependence to: ✅ Multi-model architectures ✅ BYOK ✅ Model gateways ✅ Fallback models ✅ Portable AI agents ✅ Provider-independent workflows 💡 My takeaway: The next competitive advantage in enterprise AI may not be choosing the smartest model. It may be building a system that can survive when models, vendors or contracts change. The Cursor–OpenAI situation is an early signal of a bigger transition: The AI race is moving from models to the full stack. And the companies that control more layers of that stack may control more of the future. What would you choose for your AI stack: 🔵 A tightly integrated ecosystem or 🟣 A multi-model, provider-neutral architecture? 👇 Curious to hear how AI builders and CTOs are thinking about this. #AI #ArtificialIntelligence #OpenAI #Cursor #SpaceX #Anthropic #AIAgents #EnterpriseAI #AIInfrastructure #DeveloperTools #TechStrategy

DOV'È CHE CORRONO TUTTI? Negli ultimi anni, di fronte a una buona parte della narrativa linkediniana, provo un crescente senso di straniamento. Osservo questa caotica corsa tecnologica verso non si è ancora capito dove. Escludendo l'intento speculativo di chi già prima aveva fatto lo stesso con NFC (e leggasi "qualunque novità tecnologica"), non riesco a comprendere tutti gli altri. Perché bisogna correre a prescindere dalla direzione e dalle proprie esigenze? Io ragiono, per logica, al contrario. In base alla direzione e alle mie possibilità decido l'andatura e l'attrezzatura. Se fare sponsorizzate su Meta non serve (o non ho budget), non le faccio. Se usare l'IA per fare lead generation non migliora i miei processi attuali, non la uso. Ma è un pensiero solitario. Perciò rimango a osservare, fuori dalla mia piccola bottega digitale, cosa succede. Un po' per sincera curiosità, un po' per comprendere se, forse, non mi stia perdendo qualche pezzo. -- Secondo il buon Claude, "Se invece preferisci un tono da riflessione condivisa piuttosto che da post-esca-commenti, va bene così com'è". Bello il termine "post-esca-commenti".

Nas últimas semanas, quem mandou mensagem no site do Cortex Summit perguntando sobre o evento, provavelmente falou com a ValérIA. Ela é o agente conversacional (inspirada na chefa Valéria Duarte) que construí pra tirar dúvida de quem quer saber mais sobre o evento: programação, ingresso, local, o que esperar de cada trilha. Conversa pelo WhatsApp, porque é onde as pessoas já estavam perguntando de qualquer jeito, só que espalhado entre e-mail, direct e mensagem pro time de vendas. Juntar isso num canal só, com resposta na hora, já era metade do problema resolvido. Uso a palavra "agente" com cuidado, porque virou moda chamar qualquer automação de agente, e isso confunde mais do que ajuda. A ValérIA não decide nada de negócio, não substitui ninguém do time, não tem autonomia pra prometer o que não pode cumprir. O trabalho dela é estreito de propósito: informação de baixo risco, em volume alto, sobre um domínio de conhecimento fechado (mas vivo), o evento, e só o evento. É esse tipo de escopo limitado que faz sentido automatizar sem alguém revisando cada resposta antes de sair. Automação de decisão de negócio pede supervisão constante. Automação de "onde fica o teatro" não pede. Resultado até agora: mais de 1.000 atendimentos só no período de pré-evento, sem contar o volume que ainda vem durante o Summit em si e no pós. Só que conversa individual resolve quem procura a ValérIA. Pra avisar mais de 1.300 pessoas sobre confirmação, lembrete e atualização, eu precisava de outra coisa: alcance em massa, não conversa 1:1. Daí o disparador de WhatsApp que construí em seguida. Tinha duas opções: contratar uma ferramenta pronta, ou construir uma. Construí com o Claude e ajuda do Rafael Rossini. Não foi teimosia. Uma ferramenta pronta resolveria o disparo, mas do jeito genérico dela, sem integrar com o que a ValérIA já sabia sobre cada lead, dependendo de outro contrato, outro fluxo de aprovação pra qualquer ajuste. Com IA, deu pra montar um disparador próprio rápido o bastante pra valer o esforço de construir em vez de comprar, ainda será reaproveitado. Cheguei a cogitar colocar isso tudo dentro da própria ValérIA, com gatilho de comportamento disparando mensagem automática. Tecnicamente dava. Optei por não fazer: forçar esse tipo de automação dentro de um agente conversacional ia virar gambiarra, não ferramenta própria, e eu não tinha, numa semana de construção corrida, um caso de uso claro o bastante pra justificar essa complexidade extra. O motivo mais importante de separar as duas coisas, porém, não foi técnico. Eu sabia que ia sair de licença pro nascimento da minha filha, e a operação não podia ficar concentrada em mim. O disparador nasceu vinculado à ValérIA, mas foi desenhado pra qualquer pessoa do time usar. Isso importa mais do que parece: virou canal novo de conversa com lead, não um recurso de evento que expira depois do dia 1º, e não depende de eu estar por perto pra continuar funcionando.

🎨 **DESIGN EXPERT | $70–$80/HOUR | FULLY REMOTE** Are you a **Senior Product Designer, UX/UI Designer, Design Lead, or AI-native engineer** with 5+ years of experience? 🚀 Mercor is hiring **Design Experts** to help build evaluation tasks for next-generation AI systems working across **product design, rapid prototyping, design systems, and AI-assisted development**. 💰 **Pay:** $70–$80/hour 🏠 **Work:** Fully Remote 💼 **Type:** Hourly Contract 🎯 **Experience:** 5+ years preferred 🤖 **Focus:** Product Design + AI-Assisted Development ### 🔥 TWO EXCITING TRACKS **1️⃣ Vibecoding / AI-Assisted Development** Work around: 💻 Natural-language-to-code workflows ⚡ Rapid prototyping & iteration 🤖 AI coding assistants 🛠️ Cursor, Claude Code, v0, Replit 🔄 Human-AI collaborative development **2️⃣ Product Design** Work across: 🎨 UX/UI Design 🧩 Design Systems 📐 Component Libraries 👥 User Research 🧪 Usability Testing 🚀 Rapid Prototyping 📱 Product Experience Tools may include **Figma, Framer, AI coding assistants, and design-system tooling**. ### 🧠 WHAT YOU'LL DO You'll help create challenging evaluation tasks that capture how **senior product builders and designers actually think**. ✅ Construct realistic product-design & development scenarios ✅ Build rapid-prototyping and iteration tasks ✅ Develop design-system scenarios ✅ Create reference prototypes, design specifications, and code artifacts ### 👤 WHO SHOULD APPLY? Ideal candidates include: 🔹 Senior Product Designers 🔹 UX/UI Designers 🔹 Design Leads 🔹 Design System Specialists 🔹 Product Design Managers 🔹 AI-Native Engineers 🔹 Vibecoding Experts 🔹 Rapid Prototyping Specialists 🔹 Product Builders 🔹 Startup / Independent Product Designers ### 🎯 IDEAL QUALIFICATIONS ✔️ 5+ years in product design, design leadership, or AI-native development ✔️ Experience at a product/design-focused company or strong independent/startup background ✔️ Experience owning design systems or AI-assisted development workflows ✔️ Strong understanding of user-centered design ⭐ Prior experience creating **rubrics, training data, design critiques, or evaluation frameworks** is a plus. 🤖 **This is an opportunity to use your professional design expertise to help shape how AI systems build products and understand design.** 💳 Weekly payments via **Stripe or Wise**. ⚠️ Independent contractor opportunity. Mercor notes that **H-1B and STEM OPT candidates are not supported**. 👉 **APPLY HERE:** https://t.mercor.com/cxUa9 📢 Know a senior product designer, UX/UI expert, or AI-native builder? **Tag them or share this opportunity!** #DesignExpert #ProductDesigner #UXDesigner #UIDesigner #ProductDesign #DesignSystems #UXUI #Figma #RapidPrototyping #Vibecoding #AIDevelopment #AIJobs #AITraining #AIEvaluator #AIResearch #RemoteJobs #DesignJobs #RemoteDesignJobs #ContractJobs #FreelanceJobs #Mercor #HiringNow

My AI signed a $50 billion hedge fund. The entire research and outreach process took 2 hours. Most cold outreach starts with a lead list. Ours started by teaching the AI how hedge fund data buyers actually think. We had 18 years of human-labeled bullish and bearish conviction data. Valuable data, but I had never sold into this market and did not speak the language. So the AI did 5 things: - Mapped the possible buyers: hedge funds, AI labs, data marketplaces, and sports analysts - Mined Reddit, specialist forums, research papers, and public conversations for the exact language each group used - Identified which terms signaled value to hedge fund analysts - Enriched every account and found the people who could approve a data purchase - Scored the buyers and wrote only to the strongest matches The difference showed up in one sentence. Generic outreach: "We have a unique market sentiment dataset." What we sent: "18 years of human-labeled bullish and bearish conviction, orthogonal to standard NLP-derived factors." Same data. One sounds like an outsider asking for a call. The other sounds like someone who understands how the buyer evaluates alpha. Perplexity and Claude handled the market research. Apollo and Serper found and enriched the decision-makers. The AI turned that research into the final message. 2 hours after entering a market I knew nothing about, the reply landed. He asked me to move the conversation to his work email. That became a signed client. Most companies use AI to send more messages. The better use is to understand the market so deeply that the right message becomes obvious. Volume gets you ignored faster. Buyer intelligence gets you into the room.

𝐖𝐞’𝐫𝐞 𝐇𝐢𝐫𝐢𝐧𝐠: Agentic Software Engineer III @ Deloitte (Hybrid) Deloitte’s Customer team helps organizations build stronger customer relationships through advanced analytics, Generative AI, digital products, and transformative technologies. In this role, you’ll support AI-assisted software delivery across the full development lifecycle, translating requirements into structured workflows and ensuring high-quality technology solutions. 𝐑𝐨𝐥𝐞 𝐎𝐯𝐞𝐫𝐯𝐢𝐞𝐰: Type: Full-Time Level: Manager Location: Hybrid Salary: $107,600 – $198,400 𝐑𝐞𝐬𝐩𝐨𝐧𝐬𝐢𝐛𝐢𝐥𝐢𝐭𝐢𝐞𝐬: Translate business and technical requirements into structured tasks for AI-assisted development workflows. Execute software development activities across build, testing, release, and validation. Review AI-generated code, requirements, tests, and technical documentation for quality and accuracy. Collaborate with architecture, engineering, QA, product, UX, and documentation teams. Apply engineering standards, quality practices, and delivery controls. Lead projects or workstreams while managing multiple priorities. 𝐐𝐮𝐚𝐥𝐢𝐟𝐢𝐜𝐚𝐭𝐢𝐨𝐧𝐬: Bachelor’s degree in Computer Science, Engineering, Information Systems, Data Science, Mathematics, or a related technical field. 4+ years of experience in software engineering, DevOps, SRE, quality engineering, or SDLC delivery. Experience with Python, JavaScript, Java, or another programming language. 1+ years using AI-assisted development tools such as Claude, GitHub Copilot, or Cursor. Experience reviewing code, tests, or technical deliverables for quality. Agile development experience, including sprints and backlog refinement. Strong communication, organization, leadership, and problem-solving skills. Ability to travel up to 50%. Must be authorized to work in the U.S. without future sponsorship. 𝐁𝐨𝐧𝐮𝐬: Cloud experience with AWS, Azure, or GCP. CI/CD, DevSecOps, automation, or platform engineering experience. Prompt engineering or context design for AI-assisted development. Automated testing, static analysis, monitoring, or release validation experience. Containerization and deployment automation experience. 𝐖𝐡𝐲 𝐉𝐨𝐢𝐧: You’ll work at the intersection of software engineering and agentic AI, helping shape how AI-assisted development is applied across complex customer and public-sector environments. This is an opportunity to combine hands-on engineering with leadership and emerging AI technologies. 𝐀𝐩𝐩𝐥𝐲 𝐡𝐞𝐫𝐞: https://lnkd.in/e98Qef8e

🚀 Day 29 of #60DayClaudeChallenge 🚨 Today I built Operation Lifeline: Supply Chain Crisis Lab—a full enterprise business simulation, entirely with Claude. You step in as a supply chain leader at a randomly generated company, and within seconds a crisis hits—a factory fire, a port strike, a cyberattack, a supplier bankruptcy. No two runs are the same. From there you're making real tradeoffs: 🔥 War Room—pick 3 of 6 crisis responses, and watch Cost, Inventory, Profit, Speed & Customer Satisfaction shift live 🤝 Supplier Negotiation — 4 rounds of branching decisions that shape Trust, Price & Lead Time 🏛️ CEO Boardroom — 5 leadership calls under pressure 🤖 AI Strategy — invest in 2 of 5 AI capabilities to build long-term resilience At the end, you get a full Crisis Score (0–100) across Leadership, Negotiation, Resilience, Cost Control, Risk Management, and Customer Satisfaction—plus your biggest mistake, best decision, and personalized lessons learned. This is what I love about building with Claude — it's not just "generate some code." It's designing an entire decision-making experience: the game logic, the UX, and the plain-language explanations of why each choice matters, all in one shot. Swipe through the slides to see how it works 👉 Anthropic ABTalksOnAI Anil Bajpai #60DayClaudeChallenge #Claude #AnthropicAI #SupplyChainManagement #SupplyChain #BusinessSimulation #AIinBusiness #EnterpriseAI #GenerativeAI #BuildInPublic #ArtificialIntelligence #LeadershipDevelopment #DigitalTransformation #FutureOfWork #ProductManagement #TechInnovation #AIforBusiness #LearningInPublic #Logistics #Procurement

Most Coaches Haven’t Seen This A lot of coaches are using Claude. But honestly, most are using it for one thing: “Write me a caption.” And that's where they're leaving a lot on the table. When you combine Claude + Systeme, you can speed up some of the most time-consuming parts of your business. 1. Build your marketing faster Give Claude your offer, audience research, testimonials and past content. It can help you find better hooks, write content, improve your messaging and create email sequences. Then use Systeme to turn that work into landing pages, funnels and automated emails. 2. Find what's hurting your conversions Give Claude your funnel copy, lead feedback, sales call notes and conversion data. Ask it: “What patterns do you see?” It can help you find weak messaging, common objections and places where prospects are dropping off. Then you can use those insights to improve your Systeme funnel. 3. Build your business systems Instead of keeping everything inside your head, use Claude to help turn your knowledge into: SOPs. Client onboarding. Lead qualification. Follow-ups. Sales processes. Then build those processes inside Systeme and let the automation handle the repetitive work. BONUS: MCP This is where things get really interesting. With MCP, Claude can connect with external tools and data, opening up much more powerful workflows. So you're moving from: “Claude, write this for me.” to: “Claude, help me get this done.” That's the real opportunity. You don't need AI to replace your expertise. You need it to multiply your expertise. Claude helps you think and create faster. Systeme helps you turn that work into funnels, automations and systems. And that's a much better way to use AI in your coaching business. Thanks, Navv Neet

Campaign / Strategist Manager | Lead Gen Jay 📍 Medellín, Colombia | Poblado Office Lead Gen Jay is hiring. 🚀 We’re looking for a Campaign / Strategist Manager to join our growing team in Medellín and take ownership of cold email campaigns and client results. This isn’t a role where you simply launch campaigns and check boxes. We want someone who can think strategically, solve problems, communicate confidently with clients, and use AI + automation to get better results faster. What You’ll Do • Own and manage multiple client campaigns from strategy through execution • Optimize for deliverability, replies, meetings booked, and ROI • Build targeted prospect lists and develop compelling offers • Write and test high-converting cold email copy • Analyze performance and continuously improve campaigns • Communicate proactively with clients and keep them informed • Troubleshoot problems and find solutions quickly • Use AI and automation to improve speed, efficiency, and results What We’re Looking For • 2+ years of experience in campaign management, strategy, cold email, lead generation, or a similar role • Strong understanding of cold email, deliverability, targeting, copywriting, and campaign optimization • Excellent communication and organizational skills • Proactive, resourceful, and comfortable taking ownership • Strong attention to detail • Comfortable managing multiple clients at once • Excited about AI, automation, and finding better ways to work • Able to thrive in a fast-paced, performance-driven environment • Based in Medellín and able to work from our Poblado office Experience with Instantly, Clay, Claude Code, ChatGPT, Apollo, GoHighLevel, n8n, Zapier, or Make is a major plus. Who Will Thrive Here? Someone who doesn't wait around to be told what to do. If something isn't working, you investigate it. If a client needs something, you communicate. If there's a better way to do something, you find it. We’re looking for someone who is smart, proactive, accountable, curious, and obsessed with getting better results. Ready to Apply? --> https://lnkd.in/dmbu7Mnq

Campaign / Strategist Manager | Lead Gen Jay 📍 Medellín, Colombia | Poblado Office Lead Gen Jay is hiring. 🚀 We’re looking for a Campaign / Strategist Manager to join our growing team in Medellín and take ownership of cold email campaigns and client results. This isn’t a role where you simply launch campaigns and check boxes. We want someone who can think strategically, solve problems, communicate confidently with clients, and use AI + automation to get better results faster. What You’ll Do • Own and manage multiple client campaigns from strategy through execution • Optimize for deliverability, replies, meetings booked, and ROI • Build targeted prospect lists and develop compelling offers • Write and test high-converting cold email copy • Analyze performance and continuously improve campaigns • Communicate proactively with clients and keep them informed • Troubleshoot problems and find solutions quickly • Use AI and automation to improve speed, efficiency, and results What We’re Looking For • 2+ years of experience in campaign management, strategy, cold email, lead generation, or a similar role • Strong understanding of cold email, deliverability, targeting, copywriting, and campaign optimization • Excellent communication and organizational skills • Proactive, resourceful, and comfortable taking ownership • Strong attention to detail • Comfortable managing multiple clients at once • Excited about AI, automation, and finding better ways to work • Able to thrive in a fast-paced, performance-driven environment • Based in Medellín and able to work from our Poblado office Experience with Instantly, Clay, Claude Code, ChatGPT, Apollo, GoHighLevel, n8n, Zapier, or Make is a major plus. Who Will Thrive Here? Someone who doesn't wait around to be told what to do. If something isn't working, you investigate it. If a client needs something, you communicate. If there's a better way to do something, you find it. We’re looking for someone who is smart, proactive, accountable, curious, and obsessed with getting better results. Ready to Apply? --> https://lnkd.in/dRHxVXwg

🚀 Hiring: Growth & Software Engineering Intern — User & Company Acquisition We’re hiring 5 college students for a 1-month, fully remote internship. 🎓 Eligibility: 2nd, 3rd & 4th-year students ONLY 💰 Unpaid Internship What you’ll do: • Main Work: User Acquisition & Company Acquisition • Talk to users, companies & professionals • Generate leads and build business connections • Work on a real Software Development Project 🎁 What you’ll get: • Build connections with companies & professionals • Learn how to use Claude Code at almost no cost during and after the internship • Build your own internship project • Hands-on startup experience • Growth Certificate — if you successfully perform growth, user acquisition & company acquisition work • Software Engineering Certificate — if you successfully complete the development project 📌 Important: The main responsibility of this internship is User Acquisition & Company Acquisition. The development project is an additional opportunity for students interested in software engineering and building products. 📩 Interested? Comment “INTERESTED” below. #Hiring #Internship #GrowthIntern #SoftwareEngineering #UserAcquisition #CompanyAcquisition #BusinessDevelopment #CollegeStudents #RemoteInternship #TechInternship #StartupJobs #StudentOpportunities

🚨 Se você trabalha com vendas B2B e ainda não entendeu o que está acontecendo com MCP + IA está ficando para trás. E não estou falando de usar ChatGPT ou Claude para escrever cold emails. Estou falando de algo muito maior. A IA está começando a acessar dados B2B diretamente. Empresas. Decisores. Cargos. Tecnologias. Sinais de compra. Contatos. Enriquecimento. E tudo isso dentro do próprio fluxo de trabalho da IA. Até pouco tempo, o SDR precisava abrir: → Apollo → LinkedIn → Sales Navigator → Google → ferramenta de enriquecimento → CRM → planilha → automação → outra ferramenta... Agora imagine simplesmente pedir para um agente de IA: “Encontre 500 empresas que tenham este perfil, identifique os decisores, enriqueça os dados e me entregue os prospects com maior potencial.” A IA pode fazer isso conectada a fontes de dados através de MCPs e APIs. Isso muda o jogo. Não é mais: IA + ferramenta de prospecção. É: IA + acesso a dados + agentes + automação. E já existem empresas construindo exatamente nessa direção. O GetLeads e o MoltSets são bons exemplos dessa nova ERA da arquitetura de prospecção B2B. 👉 https://www.getleads.io/ 👉 https://moltsets.com/ O mais interessante? Estamos apenas no começo. Porque quando a IA tiver acesso a cada vez mais fontes de dados, o outbound deixa de ser uma sequência de tarefas manuais. O agente pode pesquisar. Enriquecer. Cruzar informações. Encontrar sinais. Qualificar. Priorizar. E executar ações. O SDR deixa de ser apenas um operador de ferramentas. E Passa a ser o orquestrador de agentes e dados. E isso vai mudar completamente a forma como pensamos em: Prospecção B2B. Sales Development. Lead Generation. Enrichment. Outbound. A pergunta para 2027 não será: ❌ “Qual ferramenta de prospecção você usa?” Será: 🔥 “Quais dados sua IA consegue acessar? Se você trabalha com vendas B2B, Sales Tech, RevOps ou geração de demanda: comece a estudar MCP. Porque a próxima geração do outbound não será construída apenas com listas de leads. Será construída com: DATA + AI + AGENTS. E quem entender isso antes da maioria vai ter uma vantagem enorme. O outbound está mudando. Você já percebeu?

Building AI Agents: From Design Patterns to Production https://a.co/d/0cdiq1qr "Most AI agent demos work flawlessly, until real users arrive. The gap between a compelling prototype and a dependable system is where this book begins. Framing agent development as an engineering discipline, it centers on the core loop of Perceive, Plan, Act, and Observe as the foundation for building agents that are robust, interpretable, and scalable. As a companion to Antonio Gulli’s Agentic Design Patterns, Building AI Agents carries the field’s emerging architectural vocabulary from concept into practice. Where the earlier volume defines the patterns, this book implements them: showing how to translate agent design principles into reliable, production-ready systems. Organized around four proven architectural patterns -- ReAct, Chain-of-Thought, Reflection, and Plan-and-Execute -- the book takes a hands-on, pattern-first approach. Each chapter includes working code and contributes to the development of Atlas, a unified research and coding assistant that evolves from a minimal script into a production-grade multi-agent system. Coverage spans tool integration, memory and state management, multi-agent orchestration, and system evaluation. Implementations are demonstrated across leading frameworks, including LangGraph, CrewAI, OpenAI’s Agents SDK, and Google’s ADK, with designs that generalize across major model providers such as OpenAI, Gemini, Claude, and Llama. The emphasis throughout is on transferable patterns rather than vendor-specific solutions. This book is intended for software engineers integrating agents into production systems, AI/ML practitioners moving beyond chat-based interfaces, technical leads evaluating architectural tradeoffs, and advanced students working at the frontier of applied AI. Readers should be comfortable with Python and have a foundational understanding of large language models; all other concepts are developed in context."

𝐄𝐱𝐜𝐢𝐭𝐞𝐝 𝐭𝐨 𝐬𝐡𝐚𝐫𝐞 𝐭𝐡𝐚𝐭 𝐰𝐞 𝐚𝐫𝐞 𝐡𝐢𝐫𝐢𝐧𝐠 𝐚𝐠𝐚𝐢𝐧! 🚀 We deliberately gave ourselves a settling-in period after our first hiring wave. Time to get to know the team, understand how we work together, and build the foundations before opening the next batch of roles. FORMAS.AI is an ultra-dynamic place to work. We obsess over problems: a design workflow that should be 10x better, the right orchestration of models to get to the results we want, the latest AI release and what it means for us, GTM, a persistent bug nobody wants to let go of, or simply asking why something has always been done that way. 𝐀𝐧𝐝 𝐲𝐞𝐬, 𝐰𝐞 𝐝𝐨 𝐚𝐥𝐥 𝐨𝐟 𝐭𝐡𝐞 𝐚𝐛𝐨𝐯𝐞 𝐚𝐧𝐝 𝐰𝐞 𝐦𝐨𝐯𝐞 𝐟𝐚𝐬𝐭. Our team has described FORMAS.AI as high ownership, crazy fast, high autonomy, obsessive, scrappy, experimental, and full of genuinely interesting problems to solve. Somehow, all in one combo :) A few roles my network here might find interesting: → 𝐏𝐫𝐨𝐝𝐮𝐜𝐭 𝐌𝐚𝐧𝐚𝐠𝐞𝐫 | Full-time or Fractional Strong product instincts and experience. Architecture, interior design or AEC experience is an interesting differentiator. I’m also curious about mid-career pivots, for example rare architects who have become good product people. → 𝐀𝐫𝐜𝐡𝐢𝐭𝐞𝐜𝐭 𝐢𝐧 𝐑𝐞𝐬𝐢𝐝𝐞𝐧𝐜𝐞 | Fractional Architect/interior designer + good with AI. Work directly with users on real workflows, run workshops, refine our orchestration, create learning material, test the product and influence our roadmap. → 𝐅𝐎𝐑𝐌𝐀𝐒 𝐑&𝐃 | Coming soon Computational designer + AI tinkerer + builder. You script, test, prototype and increasingly build with Claude/Codex. Design + computational thinking + AI coding is a combination we’re very interested in. Show us what you’ve made. → 𝐐𝐀 𝐌𝐚𝐧𝐚𝐠𝐞𝐫 Strong QA fundamentals, systematic thinking and attention to detail. Someone who gets a strange amount of satisfaction from finding what breaks before our users do. → 𝐒𝐨𝐟𝐭𝐰𝐚𝐫𝐞 𝐄𝐧𝐠𝐢𝐧𝐞𝐞𝐫𝐬 with Archi / Comp Design / ConTech backgrounds You understand the problems from the inside and can now build the tools to solve them. → Online Marketer → Integrated Designer → Video/ Motion Designer → Various internships One more thing about how we hire. We definitely look at experience, fundamentals and mindset. But we also look for the underdogs and unconventional profiles. People with multiple interests who have gone unusually deep in a few. Career switchers with surprisingly transferable skills. Self-taught builders who became obsessed enough to get very good. Curiosity, obsession, ownership, trajectory and the ability to learn quickly matter a lot to us. See all roles at formas.ai/careers. And if you don’t see a role that quite describe you but think you belong here anyway, write to careers@formas.ai and tell us why. Please share with your crazy talented friends. ❤️ #Hiring #FORMASAI #AEC #AI

𝗜'𝗺 𝘀𝘁𝗮𝗿𝘁𝗶𝗻𝗴 𝘁𝗼 𝗯𝗲𝗹𝗶𝗲𝘃𝗲 𝘁𝗵𝗮𝘁 𝗔𝗜 𝗽𝗿𝗼𝗱𝘂𝗰𝘁 𝗯𝘂𝗶𝗹𝗱𝗶𝗻𝗴 𝗶𝘀 𝗯𝗲𝗰𝗼𝗺𝗶𝗻𝗴 𝗮 𝗚𝗧𝗠 𝘀𝗸𝗶𝗹𝗹. Not because every GTM person needs to become a software engineer. But because the distance between: "𝗜 𝘄𝗶𝘀𝗵 𝘄𝗲 𝗵𝗮𝗱 𝗮 𝘁𝗼𝗼𝗹 𝗳𝗼𝗿 𝘁𝗵𝗶𝘀" and "𝗜 𝗯𝘂𝗶𝗹𝘁 𝗮 𝘄𝗼𝗿𝗸𝗶𝗻𝗴 𝗽𝗿𝗼𝘁𝗼𝘁𝘆𝗽𝗲" has become dramatically smaller. A sales team has a problem. You can describe the workflow. Use Claude or Codex to help structure the logic. Use a vibe-coding platform to build the interface. Connect the data. Test it with real users. And suddenly, you have something you can actually put in front of the team. It may not be production-ready. It doesn't need to be. The first goal is learning Does this solve the problem? Do people actually use it? Where does it break? What should we build next? That's the part of AI product development I'm most excited about. 𝗕𝘂𝗶𝗹𝗱 → 𝘁𝗲𝘀𝘁 → 𝗹𝗲𝗮𝗿𝗻 → 𝗶𝗺𝗽𝗿𝗼𝘃𝗲. Not: 𝗣𝗹𝗮𝗻 → 𝗱𝗶𝘀𝗰𝘂𝘀𝘀 → 𝗽𝗹𝗮𝗻 → 𝘄𝗮𝗶𝘁 𝘀𝗶𝘅 𝗺𝗼𝗻𝘁𝗵𝘀. For founders and GTM leaders: 𝗪𝗵𝗮𝘁'𝘀 𝗼𝗻𝗲 𝗶𝗻𝘁𝗲𝗿𝗻𝗮𝗹 𝘁𝗼𝗼𝗹 𝘆𝗼𝘂'𝘃𝗲 𝘄𝗶𝘀𝗵𝗲𝗱 𝗲𝘅𝗶𝘀𝘁𝗲𝗱 𝗯𝘂𝘁 𝗻𝗲𝘃𝗲𝗿 𝗵𝗮𝗱 𝘁𝗵𝗲 𝗲𝗻𝗴𝗶𝗻𝗲𝗲𝗿𝗶𝗻𝗴 𝗯𝗮𝗻𝗱𝘄𝗶𝗱𝘁𝗵 𝘁𝗼 𝗯𝘂𝗶𝗹𝗱? #GTMEngineering #AIProductDevelopment #VibeCoding #Claude #Codex #AI #SaaS

Cette semaine, j'ai contribué à AionUI, et le problème résolu raconte bien pourquoi j'aime ces projets. Le souci : l'outil supposait que tous les modèles avaient 200 000 tokens de contexte. Un fallback unique. Résultat, quand tu branchais Kimi, DeepSeek ou Qwen, l'autocompactage se déclenchait au mauvais moment, faute de connaître la vraie fenêtre du modèle. Ma PR ajoute un catalogue par modèle : chacun sa fenêtre, le compactage tombe juste. Pourquoi je contribue à ce genre de projet : ces outils font tourner nos abonnements Claude, Codex, Kimi ou Gemini sur nos propres VPS. C'est la brique qui permet de gérer un business complet avec l'IA, sans dépendre de quinze SaaS. Et quand tu utilises un outil tous les jours, réparer un bug dedans, c'est réparer ton propre atelier. C'est la logique derrière Noosphere, ma plateforme GTM open source : outbound autonome, contenu inbound multicanal, inbox unifiée LinkedIn, email et WhatsApp, le tout self-hosté. Et ça m'a donné envie de relancer LaToile, mis en pause cet été : un workbench de gestion de projet AI-native, où un Manager IA orchestre des agents à rôles fixes, avec preview live, et rien ne merge sans approbation humaine. Un abonnement, un VPS, de l'open source. C'est le stack. Les trois repos sont en commentaire.

I've just open-sourced a multiplayer AI GTM Operating system that small teams can use to run their entire marketing and sales on an AI-native system without needing to set up any database, servers or SaaS subscription. 🔹 Ideal for GTM teams of 5 or less 🔹 It has a company brain where you can add full context of your company 🔹 It becomes smarter as your team uses it (learning loop baked in) 🔹 Comes with 12 pre-built AI skills 🔹 It never sends an email or spends money or publishes content or deletes file without human approval. And you can run it by adding it to your Github private repo (free account is enough). 𝐇𝐨𝐰 𝐢𝐭 𝐰𝐨𝐫𝐤𝐬? 🔹 Just give the link of the github repo (see the pinned comment) to your AI tool (Claude code, Codex, Cursor, etc) and ask it to set it up for you on your Github account. 🔹 it runs the initial setup and interviews you for 15 min to gather all your company information, which it then automatically organises in your company brain. 🔹 Then you just give access of this repo to your team members and they can start using it. 𝐓𝐡𝐞 𝐁𝐞𝐬𝐭 𝐏𝐚𝐫𝐭 👇 🔹 No team member can change the company brain, they can only propose changes which the owner has to approve (15-30 mins every week) 🔹 Team members can build workflows, agents or prompts and propose it to be shared with the team. Once the owner approves, it becomes a shared capability across your team. 🔹 Since everyone works with the same company brain, the outputs across the team becomes consistent and on-brand, always. Check the link of the repo in the comment and try it out. 👇

I just hit 80,000 LinkedIn followers. This got me to $7M+ ARR. Here's every tool I use to grow & monetize my audience: IDEATION 1. Taplio I scroll their in-app feed to spot outperforming content in my niche. You can type in keywords like "GTM", "Claude Code" or "Outbound", then filter posts that: - went live in the last 60 days - contain these keywords - crossed 400+ likes I then study the structure, hooks & CTAs of these outliers... to inspire mine. 2. Scripe Great for surfacing additional content ideas. You can aggregate posts by creator, which makes it easy to single out their best-performing hits. COPYWRITING 3. Grammarly As a non-native English speaker, I make plenty of spelling mistakes. Their extension fixes my writing in real-time. 4. Claude Code I run Opus/Fable inside Claude Code to generate hook variations, recycle older content & try new angles. Because it holds so much context on my business, it fills the gaps without me re-explaining everything each time. 5. Conductor My unified interface to effortlessly toggle between Claude Code & Codex. Their UI looks super clean (especially compared to a terminal), thus it makes writing with AI more enjoyable. VISUALS 6. Figma / Adobe Suite My LinkedIn strategy leans heavily on visuals i.e: infographics & carousels such as this one. Done right, they can multiply a post's reach by up to 10X. Our designers build carousels and infographics using these. 7. Canva My first LinkedIn carousels were made with Canva. Still the most beginner-friendly tool for social media visuals. VIDEO 8. OpusClip It automatically turns my long-form content into short-form videos. Not as good as a professional editor... But very cheap + fast. Several AI-edited clips got me > 50K views on LinkedIn. 9. Screen Studio My go-to Mac screen recorder. Always looks good for walkthroughs/tutorials as it auto-zooms on whatever part of the screen you're working on. 10. Ezgif I use it to turn short, audioless clips into .GIF files. Quick hack: If you're planning to post a sub-15-second silent video on LinkedIn... Speed it up and convert it to a sub-10-second GIF . It loops on itself & usually outperforms video format on reach/engagement. PLANNING 11. Notion My content hub. Most of my ideas get collected, drafted and prepped there. I also use it as my content calendar. Sounds cliché, but posts I plan almost always outperform posts I think of in the morning of posting. BONUS To convert our engagement, we filter ICP-fit leads that interacted with our publications. We then reach out with individually personalized messages. This 'warm outreach' flow converts at 3X the rate cold outreach does... And it can be built with 3-4 prompts via ColdIQ's MCP. Content gathers the attention. This play turns it into revenue. What tools do you recommend to get me to 100,000 followers? 👀 P.S. Thanks Ada for making me much more handsome in this picture than in real life 😄

I built something yesterday that I would've paid someone else to build a year ago. My own GTM PORTFOLIO WEBSITE. I've used AI coding tools for GTM work before, but building and actually launching a website was completely new to me. So first, I asked ChatGPT Work to do deep research on: How can I build and host a portfolio website without spending money? I went through the research, watched a few YouTube videos… and then opened Codex and started building. And honestly, it was mostly me going: “Okay build this.” “No no, change this.” “Why is this not working?” 😭 And then came hosting, deployment, domains… which was another adventure altogether 😂 But around 5 hours later I had this. Live. And that's the part I find crazy. I didn't suddenly learn web development. I learned enough to turn an idea into something real. And I think AI is making that gap between “I wish I could build this” and “I built this” ridiculously small. Anyway, here's my little GTM corner of the internet now :) Would genuinely love to know what you think.

Stop trying to pretend to be a GTM engineer, do something to become one! 1) Most of the top companies in the world use Clay, Instantly.ai, Octave. These are the GTM jobs that pay the most. You can vibe code an outbound motion with crappy tools or you can learn to build it in a governed workspaces. Your choice, but one makes you a whole lot more marketable. 2) Do some side work or freelance for an agency. Agencies see multiple clients, set ups, ICPs, scale. Visibility gives experience. Experience lands jobs. 3) Post on Linkedin, X, anywhere you can. Imagine the resume of work you could display in public. Probably would land you a job or some really good exposure. 4) Use AI: Build in Codex, Claude Code, Cursor. Spin up stuff in 20 minutes vs 20 days. Convinces hiring execs that you can deliver meaningful results quickly. Show up to the interview with a solution. Every major tool you need has APIs and/or CLI. 5) Network, join communities, Slack, etc. Your network can open up opportunities you never could have imagined.

GTM Engineer's tech stack Been getting loads of questions about this, so thought I'd make a post. Infra: MS Azure & Google workspace Sequencer: Plusvibe. Instantly, Smartlead & Email Bison are all good options here. Agent: Codex. It's currently unparalleled for knowledge work. Data: GetLeads, Clay, Storeleads, Nationgraph When it comes to data, you want to get - 1) a lot of it cheaply 2) generous api rate limits Getleads really hit the nail on the head with this one. With their unlimited model, it felt expensive paying for Apollo credits. Storeleads for specific ecom data, and Nationgraph for signals from public sector.

Marketing is changing forever. Don't miss this. Yesterday Salesforce and Anthropic announced Claudeforce. Claudeforce brings the Salesforce CRM directly into Claude. If you don't understand why this is a big deal then you're likely not actually shipping anything with AI. Agent orchestration tools like Claude, Codex, and Cursor are now the de facto hub for doing work, and legacy SaaS products are becoming plugins. Whether you're doing product development, marketing, or GTM, once you start shipping with agent orchestration tools, you quickly realize you just want to use legacy SaaS products in your agent orchestration platform. This shift has important implications for companies, SaaS providers, and marketers: • For companies... if you don't have a solution for enabling agentic access to shared knowledge, you're behind. (hint: if your employees still have to manually upload call transcripts) • For SaaS providers... if you aren't investing in a robust plugin solution, you're behind. • For marketers... if you're still shipping the "old-fashioned way" (i.e., through legacy interfaces), you're behind. But it's not too late! Getting started is super easy, just takes a little agency. Whether you're a CEO, manager, or new graduate, go build something and you'll see what I mean! And for those who are building, I'd love to know what plugins you're loving right now!

"AI didn't fail in the middle, I never defined the end" Those were the words I said to my team when I fcked up when building our GTM Outbound System. I'm writing this post becuase 1) it is important for you to know this and 2) becuase I'm waiting my claude code limits to reset lol So here's what I mean when I said "AI didn't fail in the middle, I never defined the end" There are 1001 things claude code or your agent of preference can do, but it can't build something where you don't have a clear picture of the final output. Like no, negativo, nein. ❌ Because it won't work based on your expectations and it is so easy for the AI to drift away from your idea and build something that is: 100x more complicated without specific units of work that would enable you to act on and without any relatIon from what you said at the beginning and what the final output you may have ...That was what happened to me. I started building the system with really good context as input data, but I never define the final output, so I keep building and building, and when I was expecting the output I didn't get the result that I always wanted and it was totally different from the work I put on the project, it wasn't good and I felt like all the work I put in was worthless. So for you, who's reading this and want to learn from my mystakes. Here is what you need to do so you don't have a problem like this ever: 1. You own the input and the output, the model works the middle, on what it is genuinely good at. Do not let the AI decide the scope of the project, neither the context, you own the context, it is the most important thing at the beginning of the exercise and learning how to build great context is a skill you should have. 2. If a human does it better, a human does it. People think that AI can do everything, that's why we have AI SDR Companies "but samuel, AI SDRs can help you make more money and close more deals without needing you" shut up, it doesn't work. Outline the blocks/flows/steps of the system, ask yourself the following question: Could a human being actually do a better job in this specific step? Things like communication or creative thinking: you're 1000x better than any AI so use your brain power to do them, or delegate them to your team. When people try to automate everything, nothing works correctly and that costs a lot in the long-run. I don't even have to explain why. 3. A clear output is what makes QA possible at all. You own the outputs, you need to work backwards from the output you want based on the context/inputs you have. A cristal clear output makes 80% easier, you just need to be clear, and the only way to be clear is by understading what you're trying to solve in the first place. Samuel out, bye I leave you with a meme I did using image 2 (codex) See ya later

We're excited to announce our next speaker for import_bengaluru - AbdulMajed Raja who will be representing Bangpypers and speaking on "WTF AI Agent Harness? Let's build a Baby Codex" Popularly known as 1littlecoder, while moonlighting as an agentmaxxer, Abdul Majed works at the intersection of AI and GTM at Nebius. Abdul has been dabbling with AI since it was called Deep Learning, backed by over a decade and a half of professional experience with data, computers, coffee, and back pain. A passionate teacher of the craft, he has mentored 50+ early-stage startups in AI over the last 5 years. A strong advocate of open-source models, though not ungrateful to the closed frontier LLMs. Join us to have a great day of learning, networking and more Date - 6th Sept, 2026 Location - InMobi Bengaluru RSVP: https://lnkd.in/dZN9z49h Bengaluru Tech Week #import_bengaluru

If you're about to start a new project, here's the step almost everyone skips. A vibe-coded app is rarely broken. It's #undocumented — half-decided features, a data model nobody wrote down, a spec buried in your chat history. So you fix it by re-prompting. And re-prompting. Across a build, that back-and-forth quietly burns thousands of hours and a small fortune in tokens — paying the model to guess what you meant. Codalio-blueprint ends the guessing. Free, MIT, open source. 6 skills so far, more coming: → prd-builder — rough idea in, structured PRD out → code-to-prd — reverse-engineers the PRD from what you already built → mvp-checklist — forces the cut list so v1 actually ships → arch-evaluation — will your codebase survive your requirements? → gtm-plan — channels, pricing, launch sequence → doc-generation — backlog, API sketch, onboarding doc How to start: 1/ Install: /plugin marketplace add codalio/codalio-blueprint 2/ Open #Claude Code, #Cursor, #Codex, #Antigravity or #Gemini 3/ Point it at your idea — or at the repo you already vibe-coded No signup. No demo call. No waitlist. Want it walked through live? Free online #workshop: step-by-step setup, what it costs to run (the plugin is free — you only pay your agent's tokens), a free Codalio account with starter credits, and $100 in credits for the 10 best projects built with it. Registration in the first comment — #Discord invite too, new skills drop there first. #PRDtoCode #VibeCoding #BuildInPublic

Tu sitio puede cambiar mientras tu medición queda atrapada en la versión anterior ⚡ Una migración o un rediseño puede modificar URLs, formularios, botones y rutas de navegación. El sitio continúa funcionando. Pero Google Tag Manager puede seguir buscando elementos que dejaron de existir hace meses. El problema es que estas fallas rara vez generan una alerta visible. Con la IA y las APIs, el trabajo cambia. 𝗖𝗹𝗮𝘂𝗱𝗲 𝗖𝗼𝗱𝗲 / 𝗖𝗢𝗗𝗘𝗫 pueden conectar el sitio actual, GTM, GA4 y Google Ads para: ↳ Comparar lo que mide el contenedor con lo que realmente existe en el sitio ↳ Detectar URLs obsoletas, triggers frágiles y etiquetas huérfanas ↳ Reconstruir funnels y eventos sobre la navegación actual ↳ Medir formularios únicamente después de un envío exitoso ↳ Preparar y verificar las nuevas conversiones antes de publicarlas El agente inspecciona, cruza información y construye el plan de implementación. Las decisiones sobre qué medir y cuándo publicar siguen siendo humanas. No se trata únicamente de auditar un contenedor más rápido. Se trata de evitar que las campañas optimicen durante semanas con señales incompletas sin que nadie lo advierta. Eso es 𝗪𝗼𝗿𝗸𝘀𝗽𝗮𝗰𝗲 𝗩𝟱: ✅ Diagnóstico cruzado entre plataformas ✅ Medición reconstruida sobre el sitio real ✅ Señales confiables para optimizar Menos conversiones invisibles. Más datos reales para tomar decisiones. 🆕 Ya disponible en 𝗔𝗱𝗗𝗮𝘁𝗮 𝗔𝗰𝗮𝗱𝗲𝗺𝘆 · -·-·-·-·-·-·-·-·-·-·- · Tu futuro se construye con conocimiento 👉 Súmate a la evolución en AdData.Academy 🎓 #ClaudeCode #CODEX #PaidMedia #GoogleTagManager #GoogleAds #GA4 #MedicionDigital #AdDataAcademy #V5

Stellar finale to a stellar rox.com/teams launch week. 8 of the craziest things RevOps + GTM leaders are automating… from our Revenue Builders event! 1) Steve built an entire RevOps command center in Rox: one personalized operating system he can continuously edit as the business changes. - Monitors high intent signals → sends prospects personalized outbound - Delivers a targeted view of sales pipeline, out-quarter health, and areas of focus to drive across GTM - Surfaces biggest deal risks, recommended next actions, and everything else he needs to run RevOps. 2) Romain from OpenAI showed how powerful Codex computer use has gotten. We watched it complete an entire questionnaire in real time. It filled every field, clicked through every save + submit, and finishing the workflow end-to-end in a matter of seconds. 3) Bruno from XBOW built an entire GTM engine that includes... (and more) - An expansion-readiness agent that scores every customer on likelihood to expand, then builds a thesis for how that expansion should actually happen. - A community-listening agent that continuously finds people talking about XBOW across Reddit, social, and the web - A customer triage system that can take hundreds of incoming pings and surface the 5 that matter most directly in Slack 4) Jordan from Clay showed what happens when enrichment + personalization become one automated flow. A lead comes in, gets researched and enriched, and can immediately receive personalized memes, documents, and outreach without a human touching the process. The biggest theme: agents are moving way beyond helping with individual tasks + starting to run entire parts of the revenue cycle. That’s exactly what we built Rox for! Rox.com

Yesterday I asked a Head of RevOps at a 20-person AI startup what GTM stack they're running. His answer: no GTM SaaS at all. Custom backend, Postgres for storage, a durable workflow layer for orchestration, agents doing the research. Data providers are just API endpoints they swap in and out. A year ago most RevOps people would have had no idea what these things even mean, and probably most still don't. Of course there are some downsides to building out a setup like this: technical maintanence, debugging, errors, etc. But there's also a huge upside from flexibility, cost, and ability to get creative with GTM plays. Based on how well companies like Salesforce, HubSpot, and Attio are building out their headless functionality, I wouldn't recommend this for more established companies but you should 100% be using Claude/Codex to orchestrate across your tech stack and drop all the tools that don't have APIs available. They're just going to slow you down. I think we'll start seeing more AI native companies building out their tech from scratch like this. It's too fun what you can do when you can control you're entire GTM from a terminal :)

OpenAI’s expansion in Brazil made me think about a question: How should a global AI product enter a local market? The traditional playbook often looks like: Choose a market → Localize → Launch campaigns → Acquire users But Brazil suggests another path: Localize enough → Observe organic adoption → Follow product pull → Invest deeply 1️⃣ Minimum localization makes adoption possible Organic adoption does not mean localization is unnecessary. Users still need the product to be accessible and usable in their own language. But there is an important distinction: Making a global product locally usable ≠ deeply investing in a market Language localization opens the door. It doesn’t tell you how deeply to enter the market. Once the basic friction is removed, real user behavior can start revealing where demand already exists. 2️⃣ Organic adoption reveals where the product pull is Brazil was already showing strong adoption before OpenAI established its local team in São Paulo. According to OpenAI, Brazil is now: ▸ A top-three market globally for ChatGPT weekly active users ▸ Generating around 215 million messages per day ▸ The second-largest market for OpenAI API developers ▸ The largest Codex market in Latin America At this stage, the GTM question is no longer: Will Brazil want ChatGPT? It becomes: How do we deepen a market that already wants it? Organic adoption becomes more than a growth metric. It becomes a form of market research. 3️⃣ Local behavior tells you how to localize The next step is not simply translating more content or spending more on local campaigns. It is understanding how local users are actually using the product. Brazil, for example, is one of the strongest markets for ChatGPT Images and also shows high Voice usage. OpenAI’s local expansion also includes initiatives such as a São Paulo Creators Day, Portuguese-language developer programs, hackathons, education initiatives, enterprise engagement, and public-sector partnerships. This is what I find most interesting: Localization can become behavior-led. Instead of applying the same GTM playbook to every country, companies can use local usage patterns to decide which users, use cases, communities, and ecosystems deserve deeper investment. ✨ For global AI products, organic adoption can become a signal for where to invest — and local behavior a guide for how to invest. Sometimes the users choose the market first — and their behavior tells you how to enter it. #OpenAI #GlobalGTM #AIProduct #ProductMarketing

Claude code’s performance has fallen off a cliff recently, and everyone seems to be talking about it on here. But I think I know what happened… Claude Code is one of the main tools in my tech stack (probably second only to Clay) and something I use everyday. Now I haven’t been using it for software dev but mainly for gtm stuff like editing Smartlead campaigns through the API, manipulating / joining datasets, scraping the web etc. But then a couple weeks back it started going wild. I would give it a fairly basic task and it would go off on random tangents running side quests that were completely unnecessary and give long complicated outputs. Put simply: burning credits and wasting time. I was on the cusp of switching to codex until I brought it up in the StackOptimise ⚙️ weekly GTME call, and the AI wizards Done Miladinov and Muhammad Rafay came to the rescue. “Yeah Opus 5 is sh*t. Switch back to 4.8” I hadn’t even realized the model had changed. Lesson learnt. So I did and the Claude I know and love returned. Anyone else noticed this lately?

We´re still looking for Senior GTMEs based in Pakistan, you can make $5k-$6k if you have: - At least 3 years agency experience - Built, managed and scaled intent driven Outbound. - You MUST very good at Clay, extra points if you are using Claude Code or Codex to run your daily GTM ops. If you don´t have any of the above, please don´t apply. There are no openings for juniors atm. Application link is in the first comment.

At Rence (YC F26), everyone from GTM to engineering contributes to the codebase weekly. A year ago this would be reckless. Today it’s just the optimal play, and has been the most insane productivity unlock we’ve had so far. Speed is everything at our stage and tokens are abundant, so leaving them unused is a strictly worse decision than letting someone try something. Pre-AI, engineering-orgs happily let junior engineers ship into production, not because they were always right, but because they learn fast and someone reviews the output. A non-engineer + Fable 5 is arguably 8.75x this today (trust), and has better conditions to learn, fast. None of this works without engineers, of course. More people shipping means more review, and the fix isn’t to review harder, it's building an environment where agents are more likely to get it right the first time. Anyways, if anyone has an idea, they should be allowed to execute, and below is an image of our beloved Codex executing our team.

One viral AI prompt burned $1,700 in tokens. It's called the gauntlet loop, and it's the best prompting idea of the year. Builder agents make the work. A blind critic compares it against a real reference. Anything that loses goes back for another round. Matt Shumer used it to build a Call of Duty clone from three paragraphs of text. 55,000 lines of code, zero hand-made assets. The internet lost its mind. Then the bills arrived. $1,200 for an F1 game. $1,700 for a GTA attempt: 22 hours, 86 agents. The top Reddit comment on the whole trend: "guaranteed token burn with fingers-crossed results." The technique is brilliant. The spending model is unsustainable for most. So I wanted to put my own spin on it with spending guardrails. Same loop, one new rule: it refuses to start without a budget. You give it 50% of your usage window and it sizes the entire run to fit. Runs out? It stops, shows you what the spend bought, and asks before touching more. 50, then 75, then 95. Never on its own. I tested it with the same one-line brief, same model, run twice. The naive single pass shipped a landing page with a fake "as seen in Wired and TechCrunch" press bar. Completely invented. Nothing in a single pass ever asks "says who?" The gauntlet build couldn't get away with that. Its critic rejects any claim that can't point to evidence. Contrast ratios computed. Keyboard actually pressed e2e. Every number traceable to the brief. GTM teams, this is where it gets useful for you. Point it at a landing page, a competitor battlecard, a launch email sequence, or a case study. The critic checks marketing claims the same way it checks code: no invented stats, no fake logos, no "as seen in" that legal never approved. brief becomes the quality bar. What ships is what passed. Free and MIT licensed. Works on Claude Code, Codex, Hermes, or agent of your choice. Install is one command. Link in the first comment. Try it out, if you like it give it a star! Real question for anyone running agent fleets: do you cap your runs, or let them cook?

Our SaaS had a massive problem in august. Only 6.25% of our signups got value out of our product. Welcome to "Build OXYGEN in public" week 1, where I share our weekly learnings, f*ckups and wins. So here is where we screwed up: [1] We asked for a credit card on the trial period 50% dropped off right there. So we essentially invested time, money and energy to acquire a user and 50% of them just droped off on a pay wall. [2] Activation was 12.5% Of the half that got past the credit card wall, 1 in 8 reached value (which we measure in wether the user has set up a workflow, sequence or table in our product). 50% x 12.5% = 6.25%. 1 in 16 signups. [3] Churn was 13% 13% monthly churn means we replace our entire customer base every 8 months. It's well known that GTM tech has high churn, but this is way too high. [4] Support was a disaster We launched and people texted us on LinkedIn, Slack and WhatsApp. Tickets ran through a vibecoded system in our CLI and MCP 🫠 So here what we shipped to solve all of that: [1] A free tier, no credit card Plus 10$ in credits gifted on every signup. So you can find out if it works before you pay us anything. [2] A simpler onboarding Less steps between signup and your first table, sequence, workflow. We also invested time into creating some templates which can be used as a base. [3] An in app AI copilot It builds the setup for you if you don't want to open a terminal. (I still believe everybody should run this through Claude Code or Codex, but that can't be the only way in). [4] Plain as our support system Slack, in app chat and an API based option in one inbox. Makes it much easier for us and our agents to help out our customers. So here are my goals for the next 30 days: - activation from 12.5% to 25% - churn from 12.5% to 8% I ranted a lot in this post, but there has been massive wins too! We stopped our entire marketing and sales engine to fix our leakages first, but we still grew our MRR and signups. The customers that use us, love the product and tell others about it. Now that we make our product much more intuitive we are ready to go all in on GTM again heheh. So follow Tim Scheuer to read about our progress next Friday.

Most companies use AI to create more GTM activity I think the bigger opportunity is to make every campaign teach you something that improves the next one Imagine you’ve already run 30 campaigns For each campaign, you know who you targeted, which problem you thought they had, what information you used, what you sent, and whether it produced meetings or revenue The problem is that this information is usually spread across different tools, or simply forgotten when the campaign ends What if you kept a simple record of every campaign? Then AI could compare all of them and help you answer questions like: Which problems actually led to revenue? Which information helped us choose the right companies? Where did the results match the ROI we promised? Which campaigns looked good but produced nothing? The ROI part matters because a prospect should understand what they could get from working with you and why you believe that result is possible Suppose you sell software that helps finance teams finish their monthly reporting faster “AI for finance teams” doesn’t say much A better message could explain how much time similar companies saved, what slowed them down before, and which details you used to estimate the result for this specific company You’re probably asking, how do you build this? Start with 10 campaigns that worked and 10 that didn’t Write down the audience, the problem, the evidence you used, the result you promised, and what actually happened Give that information to Claude/Codex and ask it to find the patterns, then review those patterns yourself and decide which ones are worth testing again Your next campaign now begins with everything you learned from the previous ones Run the campaign, keep what you learned, and use it the next time After enough campaigns, you have your own record of what actually produces revenue, and that becomes very difficult for another company to copy

GTM Engineers - your in real danger of falling behind if you don't evolve quickly: A few weeks ago, Clay released Workflows. In one swoop, they changed GTM orchestration forever. Think back a few years ago when Clay released tables. They took CSV files, and brought them to life in the Cloud. Microsoft, in all its billions, didn't figure out how to activate a table and a row, a concept they invented, yet Clay did. Now here we are in 2026. More people use Claude, OpenAI, Cursor, etc. All Clay's customers. So what did Clay do, build a managed layer for agents. They entered Phase 2 of GTM dominance. In doing so, they found the gap that was missing in GTM. No limit executions, python code (think cleaning, classification, etc), opened up the API to build on top of it for customers. I'm not joking when I say they swallowed the 5 year strategy of Zapier, N8N, and Make in one swoop. How do I know, because I build stuff quick. Something new comes along, I try it. There is nothing I can't build now using workflows. I'll never build a table again. Save this post, because in a year, you will look back and think, glad I listened. Governed GTM at scale, from the AI of you choice has arrived, the only question will be, did you miss it. PS: pardon my Codex pet, he refused to leave the image:)

We’re hiring across the teams helping startups build, scale, and tell their stories with OpenAI. These are truly once-in-a-career opportunities working with the sharpest minds in AI (only a little biased)! Sharing a few roles that will work on/closely with our team & have an outsized impact in how we show up for builders around the world: Startup Content Marketing Manager (SF): Get close to founders building with OpenAI, uncover the stories and lessons behind their products, then turn them into high signal content that builders actually want to read, watch, and share. 👉🏼 https://lnkd.in/gVVcb_nN Integrated Marketing Manager, Developers & Startups (SF): Bring large-scale brand campaigns to life, shaping the strategy, briefing agencies, landing the creative, and making sure builders are at the center of the story. 👉🏼 https://lnkd.in/gQJnX-Hw Growth Programs Lead, Startups (SF): Build the operating system that helps startups grow with OpenAI by turning credits, partnerships, and GTM motions into scalable, measurable programs. 👉🏼 https://lnkd.in/gJbkM2JA Account Director, Startups (Global): Be a trusted partner to founders. Understand what they’re building, anticipate what they need next, and help them get more value from OpenAI at every stage of their journey. (Fun fact, 80% of our Startups GTM team are former founders!) 👉🏼 https://lnkd.in/gGmmCjPk Resource Manager, Business Marketing (SF or Remote): Help our team move faster and smarter by matching the right work to the right people, simplifying how projects get prioritized, and building AI-powered workflows (we love Codex). 👉🏼 https://lnkd.in/g23-J5Cj // Please note I won't be able to respond to DMs but recommend applying directly as roles are filled quickly. All open roles can be found at openai.com/careers.

The best way to learn GTM Engineering? Build something you actually need. I've realized that reading about AI agents, automation, and GTM workflows is useful. But nothing compares to trying to build one. You quickly discover: → Your data isn't as clean as you thought. → Your workflow has more edge cases than expected. → Your prompt isn't as reliable as you expected. → The automation breaks somewhere you didn't anticipate. And that's exactly where the learning happens. That's why over the next few months, I'm trying to spend less time asking: "What can this tool do?" And more time asking: "What can I build with it?" Claude. Codex. n8n. Vibe-coding platforms. AI agents. I'll be experimenting with all of them to build practical GTM systems and small products. Some will work. Some definitely won't. 😄 I'll share both. Build → break → learn → rebuild. That's the journey. 💬 For everyone building with AI: What's the most useful thing you've built for yourself recently? #GTMEngineering #AI #GTM #AIAgents #Automation #VibeCoding #RevOps

Team-led content is cheat code for B2B growth in 2026. Companies are generating millions with exactly this strategy ↓ (Save it before it gets lost in the feed.) Most companies still create content like this: Open a blank AI chat. Write a prompt. Fix the generic draft. Repeat everything tomorrow. Here is how best teams operate. They create a folder which claude or codex can access with: 𝟭. 𝗕𝘂𝘀𝗶𝗻𝗲𝘀𝘀 𝗯𝗿𝗮𝗶𝗻 ICP, positioning, messaging, offers, proof, objections, customer language, and the founder's point of view. 𝟮. 𝗥𝗲𝗳𝗲𝗿𝗲𝗻𝗰𝗲 𝗹𝗶𝗯𝗿𝗮𝗿𝘆 The best posts, hooks, infographics, carousels, and images are saved as quality benchmarks. The system studies the patterns without copying the work. 𝟯. 𝗣𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻 𝗽𝗹𝗮𝘆𝗯𝗼𝗼𝗸𝘀 Step-by-step instructions for posts, carousels, infographics, reviews, and repurposing. HTML templates keep every visual consistent and editable. 𝟰. 𝗜𝗱𝗲𝗮 𝗲𝗻𝗴𝗶𝗻𝗲 Reddit threads reveal pain. X shows emerging conversations. Competitors reveal crowded angles. Sales calls provide buyer language, objections, and proof. 𝟱. 𝗥𝗲𝘃𝗶𝗲𝘄 𝗹𝗼𝗼𝗽 Every draft is checked against the ICP, positioning, voice, proof, hook quality, and pipeline outcome before it ships. Now a team member does not need to guess what the founder means. They open the workspace, follow the playbook, and improve the shared system with every post. AI handles execution. The team's accumulated context protects the quality. What is your content creation workflow with AI? ♻️ Repost to help your founder friend nail GTM ➕ Follow Kanchan Bhatta for GTM systems My Tool Stack: Claude Code Cursor Google Workspace Reddit, Inc. X Higgsfield AI #GTM #B2BSaaS #FounderLedGrowth

GTM skills are evolving fast. Yesterday, I got to build one. At the GTM Skillathon in Bucharest, together with Lidia Mititelu , we built a skill focused on AI Search Optimization and online visibility. For our case study, we chose SalesOMMO. The goal was simple: understand where they currently stand online, who they compete with, what search opportunities exist and what they could improve. Because we only had around 2 minutes of execution time, we had to be very intentional about the workflow. Instead of making the agent search for everything from scratch every time, we first used Apify to extract the relevant web and search data, then fed that context directly into the agent before running the analysis. That made the workflow faster and gave the agent better context from the start. From there, the skill could compare the business with its competitors and turn the data into clear recommendations for improving online visibility. That was probably my favorite part of the challenge, not just building something with AI, but figuring out how to make it actually useful under real constraints. And honestly, the community was just as important as the project. A room full of people building, testing ideas, helping each other and shipping in a few hours is exactly why communities like Builders House are so valuable. Huge thanks to Alexandru C. ,Formidable Builders, Builders House, OpenAI and Apify for making it happen. And congratulations to all the winners, amazing gtm skills !🥇 And, of course, thanks to Lidia Mititelu for building this with me. Github repo: https://lnkd.in/d5DSYTVH Fun fact: we pushed our last commit 30 seconds before stop coding… so naturally, we only managed to take one photo together 😂 #BuildersHouse #FormidableBuilders #AgentSkills #AI #Codex #OpenAI #Apify #GTM #SEO #BuildInPublic

We had the best time yesterday at GTM Skillathon. ⚡️ Together with the one & only Radu Minea, I enjoyed every moment building a skill for OpenAI Codex. And the best part was that we met so many talented & insightful people, you can check out all the projects in the public repo. Thank you, Formidable Builders, Builders House & everyone involved. See you soon! 👀

GTM Engineers, you have to understand how everything works, its your job! By now, we all heard of ChatGPT, Gemini, Claude. But have you used Codex, Claude Code, or Cursor. Did you know Cursor is owned by SpaceX and because of that OpenAI is leaving in November. Do you understand Open Source, have you heard of GLM, are you familiar with open weights. Can you install a CLI or MCP? Its a lot, I know. But as a GTM, you are expected now to be just as much an AI Engineer at you are a GTM Engineer. Are you testing Grok and Devin. Did you know Grok has its own computer, yet most work you do in Codex is on your local machine. Your boss will expect you to know how data will be protected, what is being shared and most importantly, what does everything cost. You need to be able to explain why you need what you need to do your job. Your business has a budget, and unless you work for a unicorn, its not unlimited!

Your GTM workflows should be in github right now. maybe not a hot take, but i just got off a call where everything was either in spreadsheets or siloed in 6 different tools. makes no sense. grab your API keys. build a knowledge base with claude, codex, whatever agent you use. version control it. then you just ask: what’s my deal state? what were the reply rates on the last campaign? no digging through 500 tabs. one answer, with evidence, that tells you how to make the next campaign better. Drake gets it 🤣

🚀 AI isn’t just ChatGPT — it’s an entire ecosystem. Most people start with ChatGPT for everything, and that’s perfectly fine. But as your AI skills grow, the real advantage comes from combining specialized tools for specialized jobs. 🟢 Beginner 🤖 ChatGPT — writing, research, brainstorming, analysis & everyday AI tasks 🔵 Intermediate 🧠 Claude — deeper thinking & long-form work 🎙️ ElevenLabs — AI voice 🎬 Higgsfield — AI video 📱 Emergent — app building 🎞️ Opus Clip — video repurposing & animations 💬 ManyChat — DM & marketing automation ⚡ Advanced 💻 Claude Code — AI-powered development from the terminal 🎯 Clay — lead generation & enrichment 🕷️ Apify — web scraping & automation 🔗 MCP — connecting AI with tools and data 🎥 Artlist — creative video production 🤖 Codex — coding & task automation 💳 Stripe — payments & monetization 🌐 Manus — websites and autonomous workflows The real AI skill in 2026 isn't knowing one AI tool. It's knowing which tool to use, when to use it, and how to connect them together. From prompting → creating → automating → building → monetizing, AI becomes much more powerful as you move deeper below the surface. 💡 Don’t just learn AI. Build an AI stack that works for you. Which level are you currently at — Beginner, Intermediate or Advanced? 👇 #ArtificialIntelligence #AI #ChatGPT #ClaudeAI #AITools #GenerativeAI #Automation #ClaudeCode #Codex #MCP #AIWorkflow #AIAutomation #FutureOfWork #Productivity #TechTrends #DigitalTransformation

Can we actually rely on the results AI brings? My article👉 https://lnkd.in/gUXQ53Te If you're fantasizing about building an agent team so you can sit back and watch users and revenue grow on their own, you've got it very wrong. I know there are a lot of people posting right now about how much money AI made them, or how many followers AI got them for them — either they're lying, or they spent an enormous amount of time training that AI to do the work. My take: AI can only speed up how fast you find the path that works. The call is still yours. Last week I sat in on an online session from Zapier and Clay about how they grow through social media. In the Q&A, someone asked how they use AI these days. Almost every one of them said they don't really depend on AI. They use it to dramatically speed things up, or to visualize results really well. But the core decisions aren't made by AI. For senior marketing, you have to keep your feel for it. You can't let your AI understand your prospects better than you do — if it does, the role should go to the AI. Your customers' pain and needs — you have to be able to say them out loud, not have to look at some long report AI hands you every time. You need to sit inside the customer quotes AI feeds you and feel what they're feeling. Feel is really just empathy. That's something only humans have. An example: I'll have Codex browse Reddit for me and pull together what my prospects are saying is painful, and Codex hands it back to me organized. But what I actually reply with, I have to think through myself. That's how you actually feel the pain your customer is in — and then you can have a real conversation with them. Customers aren't going to pay because of the AI video or image you made This is something I figured out recently. I'm sure when you scroll short videos, you've come across cute little cats and dogs dancing — very short clips that somehow have tens of thousands of likes. You probably liked them too, and were amazed AI can make videos like that now. But look, the audience for that kind of video is not your customer. You all know a B2B customer's decision cycle is long — they usually have to read your product materials, compare you against competitors, run a trial, and discuss internally before they commit. You might generate a product intro demo or an animation with AI, but the customer is absolutely not going to buy your product because of that video. Making people laugh with an AI video is easy. Making a B-end customer pay for something? Not likely. You can't expect someone to see a fake thing and then hand you real money for it. There's still a gap between what AI delivers and what a product actually delivers.

Hay una diferencia enorme entre usar IA y construir un sistema de trabajo alrededor de ella. Para mucha gente, IA sigue significando: → abrir ChatGPT → escribir un prompt → copiar la respuesta Y eso es solo la punta del iceberg. Cuando empiezas a especializar las herramientas, el panorama cambia: Claude para determinados procesos de razonamiento. Gemini para trabajar con grandes cantidades de contexto. NotebookLM para documentación. ElevenLabs para voz. Claude Code y Codex para desarrollo. Clay + Apify para datos y prospección. Supabase para infraestructura. Metricool para analítica. Y después aparece una capa todavía más interesante: conectar todo mediante agentes, APIs, MCP y automatizaciones. El objetivo no debería ser aprender 50 herramientas. Debería ser construir un stack en el que cada herramienta haga aquello para lo que realmente es buena. La pregunta interesante es: ¿Sigues utilizando una IA para todo o ya has empezado a construir tu propio stack? #IA #InteligenciaArtificial #Automatizacion #Productividad #Tecnologia

AI gets more useful when you stop treating every tool the same. Some tools help you think. Some help you create. And some help you build entire systems. That is why I like looking at AI in 3 levels. Beginner Start with tools that improve everyday work: → ChatGPT for general problem-solving → Claude for deeper thinking → Perplexity for research → Notion AI for notes → Grammarly for writing → Gamma for presentations → Canva AI for design At this stage, the goal is simple: save time and improve output quality. Intermediate Now move from assistance to production: → ElevenLabs for voice → Higgsfield for video → Opus Clip for repurposing → ManyChat for DMs → HeyGen for avatars → Zapier for automation → Emergent for building apps You are no longer just asking AI questions. You are using it to create repeatable workflows. Advanced This is where things become much more powerful: → Claude Code for terminal workflows → Codex for coding tasks → Clay for lead generation → Apify for scraping → MCP for connecting tools → n8n for workflows → LangGraph for agents → Supabase for backend infrastructure The biggest shift is not learning more AI tools. It is learning how to connect them. That is when AI stops being a collection of apps and starts becoming infrastructure for how you work. Want to learn how to turn LinkedIn into a predictable lead generation engine? Join my Skool community - https://lnkd.in/gMRGNTv9

Stop trying to pretend to be a GTM engineer, do something to become one! 1) Most of the top companies in the world use Clay, Instantly.ai, Octave. These are the GTM jobs that pay the most. You can vibe code an outbound motion with crappy tools or you can learn to build it in a governed workspaces. Your choice, but one makes you a whole lot more marketable. 2) Do some side work or freelance for an agency. Agencies see multiple clients, set ups, ICPs, scale. Visibility gives experience. Experience lands jobs. 3) Post on Linkedin, X, anywhere you can. Imagine the resume of work you could display in public. Probably would land you a job or some really good exposure. 4) Use AI: Build in Codex, Claude Code, Cursor. Spin up stuff in 20 minutes vs 20 days. Convinces hiring execs that you can deliver meaningful results quickly. Show up to the interview with a solution. Every major tool you need has APIs and/or CLI. 5) Network, join communities, Slack, etc. Your network can open up opportunities you never could have imagined.

I've been seeing the reintroduction trend on LinkedIn for a while now. Two announcements in one month, the Bootprint 🥾 (prev Clay Bootcamp) rename and the n8n one, and I figured most of you only know me from the build posts. So I'm finally giving it a shot. → I'm Anna Bui. I live in Da Nang, Vietnam, and I've been working remotely since 2020. → I grew up in Hanoi, the capital. Then I fell in love with the sea and beach life, and that's how I ended up here. → I did not start in tech. My first job was leading an Xbox customer success team. After that: Office 365 tickets, customer service for a gaming chair brand in Singapore, booking podcast guests, then KYC and fraud at a crypto exchange. → Automation found me while I was a project manager. I started building workflows with Zapier and ClickUp, and the love started there. (Yes, Zapier. Everyone starts somewhere.) → n8n came next. Yesterday's post covers that one. → Then I got curious: Claude Code, Trigger.dev, Cloudflare Workers, Supabase. I talk to Claude and Codex out loud most days. → I have a Bengal cat named Quebec. A Bengali man living in Vietnam, named after a Canadian city, and I speak to him in English. Mr Worldwide. → Outside work I'm either in the water or on a bike. Swimming nonstop, cycling the rest of the time. → Just over a year ago I was job hunting with no idea what I was looking for. Mark Colgan introduced me to Clay Bootcamp, and one call with Nathan Lippi 🥾 and Patricia Chu 🥾 later I was on the team. → Today I coach at Bootprint 🥾 (prev Clay Bootcamp), and as of this week I'm an n8n ambassador for Vietnam, building the community from Da Nang. First event soon. Both jobs are the same job: sit with someone who's convinced they can't build it, until they can. That's me beyond the build posts and arrow lists. Now tell me one thing I couldn't guess about you.

GTM Engineer's tech stack Been getting loads of questions about this, so thought I'd make a post. Infra: MS Azure & Google workspace Sequencer: Plusvibe. Instantly, Smartlead & Email Bison are all good options here. Agent: Codex. It's currently unparalleled for knowledge work. Data: GetLeads, Clay, Storeleads, Nationgraph When it comes to data, you want to get - 1) a lot of it cheaply 2) generous api rate limits Getleads really hit the nail on the head with this one. With their unlimited model, it felt expensive paying for Apollo credits. Storeleads for specific ecom data, and Nationgraph for signals from public sector.

Stellar finale to a stellar rox.com/teams launch week. 8 of the craziest things RevOps + GTM leaders are automating… from our Revenue Builders event! 1) Steve built an entire RevOps command center in Rox: one personalized operating system he can continuously edit as the business changes. - Monitors high intent signals → sends prospects personalized outbound - Delivers a targeted view of sales pipeline, out-quarter health, and areas of focus to drive across GTM - Surfaces biggest deal risks, recommended next actions, and everything else he needs to run RevOps. 2) Romain from OpenAI showed how powerful Codex computer use has gotten. We watched it complete an entire questionnaire in real time. It filled every field, clicked through every save + submit, and finishing the workflow end-to-end in a matter of seconds. 3) Bruno from XBOW built an entire GTM engine that includes... (and more) - An expansion-readiness agent that scores every customer on likelihood to expand, then builds a thesis for how that expansion should actually happen. - A community-listening agent that continuously finds people talking about XBOW across Reddit, social, and the web - A customer triage system that can take hundreds of incoming pings and surface the 5 that matter most directly in Slack 4) Jordan from Clay showed what happens when enrichment + personalization become one automated flow. A lead comes in, gets researched and enriched, and can immediately receive personalized memes, documents, and outreach without a human touching the process. The biggest theme: agents are moving way beyond helping with individual tasks + starting to run entire parts of the revenue cycle. That’s exactly what we built Rox for! Rox.com

This is where AI becomes much more powerful. 🚀 It’s not about using one AI tool. It’s about connecting the right tools to create an AI-powered workflow. → Claude Code — terminal & development workflows → Codex — coding & automation tasks → Clay — lead generation & enrichment → Apify — web scraping & data collection → MCP — connecting AI to tools & data → n8n — workflow automation → LangGraph — building AI agents → Supabase — backend infrastructure The biggest shift is not learning more AI tools. It’s learning how to connect them. That’s when AI stops being a collection of individual apps and starts becoming infrastructure for the way you work. The future isn’t about asking, “Which AI tool should I use?” It’s about asking: “How can I connect these tools to automate the entire workflow?” That mindset can transform productivity, operations, sales, development, and decision-making. AI is becoming a connected system—not just a collection of tools. What AI tools are you currently connecting in your workflow? 👇 #AI #GenerativeAI #Automation #AIAgents #Productivity #n8n #MCP #AItools

Claude code’s performance has fallen off a cliff recently, and everyone seems to be talking about it on here. But I think I know what happened… Claude Code is one of the main tools in my tech stack (probably second only to Clay) and something I use everyday. Now I haven’t been using it for software dev but mainly for gtm stuff like editing Smartlead campaigns through the API, manipulating / joining datasets, scraping the web etc. But then a couple weeks back it started going wild. I would give it a fairly basic task and it would go off on random tangents running side quests that were completely unnecessary and give long complicated outputs. Put simply: burning credits and wasting time. I was on the cusp of switching to codex until I brought it up in the StackOptimise ⚙️ weekly GTME call, and the AI wizards Done Miladinov and Muhammad Rafay came to the rescue. “Yeah Opus 5 is sh*t. Switch back to 4.8” I hadn’t even realized the model had changed. Lesson learnt. So I did and the Claude I know and love returned. Anyone else noticed this lately?

We´re still looking for Senior GTMEs based in Pakistan, you can make $5k-$6k if you have: - At least 3 years agency experience - Built, managed and scaled intent driven Outbound. - You MUST very good at Clay, extra points if you are using Claude Code or Codex to run your daily GTM ops. If you don´t have any of the above, please don´t apply. There are no openings for juniors atm. Application link is in the first comment.

$5B valuation. 14,000 customers. $100M+ ARR. And they just handed their most technical users the exit. Clay shipped an API. Every serious observer said they never would, and the logic was sound: an API doesn't kill revenue directly, it kills stickiness — and lost stickiness kills revenue on a delay. Three things were holding that moat: → Aggregated access to 150+ data providers → Your templates, prompts and logic living inside their environment → An agency channel selling on their behalf An API threatens two of the three. So why do it? Because the migration had already happened. The most technical users — the ones spending the most — had stopped opening the UI. They were running the same workflows from Claude Code, Cursor, Codex. The API didn't open a door. It acknowledged that people were already climbing out the window. That's the part worth sitting with. A company at that scale looked at its highest-value segment and concluded the interface was no longer where the relationship lived. Better to be the thing the agent calls than the tab nobody opens. I think that's the right read, and I think it's about to happen across the category. Within a year, the vendors that treat the agent as a first-class user — real API, MCP server, bulk operations, machine-readable errors — take the most technical and highest-spending slice. The ones that don't become the tools a human has to babysit. Babysitting is what churn looks like six months before it shows up in the numbers. Which of your vendors could an agent actually drive today?

GTM Engineers - your in real danger of falling behind if you don't evolve quickly: A few weeks ago, Clay released Workflows. In one swoop, they changed GTM orchestration forever. Think back a few years ago when Clay released tables. They took CSV files, and brought them to life in the Cloud. Microsoft, in all its billions, didn't figure out how to activate a table and a row, a concept they invented, yet Clay did. Now here we are in 2026. More people use Claude, OpenAI, Cursor, etc. All Clay's customers. So what did Clay do, build a managed layer for agents. They entered Phase 2 of GTM dominance. In doing so, they found the gap that was missing in GTM. No limit executions, python code (think cleaning, classification, etc), opened up the API to build on top of it for customers. I'm not joking when I say they swallowed the 5 year strategy of Zapier, N8N, and Make in one swoop. How do I know, because I build stuff quick. Something new comes along, I try it. There is nothing I can't build now using workflows. I'll never build a table again. Save this post, because in a year, you will look back and think, glad I listened. Governed GTM at scale, from the AI of you choice has arrived, the only question will be, did you miss it. PS: pardon my Codex pet, he refused to leave the image:)

Good news! My mobile shop is open today at The Charlottesville City Market. Thanks to the Launch Pad Program and The Community Investment Collaborative that makes this possible for collaboration. Featuring my book “The Dakini Codex.” A book that explains complex form of science and spirituality and presents them in an easy to understand format with lyrical poetry. It has been featured at the world’s largest book signing event in Frankfurt, Germany. It has been nominated for The Eric Hoffer Award and placed in the museum at The University of Science and Philosophy in Waynesboro, Virginia. My crystal jewellery are Somatically attuned and reiki infused for wearable forms of healing art. Serenity Hope is featuring two new product lines! A crystal worry stone set in a marbelized polymer clay cabochon. It can be worn as a pendant and used as a fidget to ease anxiety. I also have upscale recycling pendants from bottle caps. They are a toast to your favourite drink! A beautiful example of how we can keep mementos of precious memories while being conscious of how we live in right relation to the planet and her resources. Come and visit! We love to see faces from the community. Together we make a difference. 🕉️🙏🏻

I spent 127.33 hours on this INSANE resources. It's SO F****** good that REAL GTMes are asking why I'd give away for free. . . . (because I want everyone to eat good this christmas) This resource laid out exactly the A-Z on how AJ Cassata and I build cold email systems that get results like: - $1.2 Million in Cash Collected in 7 months - 389 leads in 4 weeks - $700,000 funded in 3 months It's exactly how AJ and I made $40M for our clients via outbound. It's our secret "Big Mac" sauce on how to: 1. Land in the primary (with Inboxkit and SendKit best tips) and expert trainings with Rahul Lakhaney 2. Build list full of people who salivates about your product with AI Ark - expert training with Hamied Elias Kermani 3. Me and AJ Cassata best, most juciest copywriting and offer training that would make Alex Hormozi and Sabri Suby proud 4. All of our claude skills we use to run outbound with tools like Clay, LocalPipe.io and more Free access below. If you want more stuff like this - follow Harvey Le 🐦‍🔥

GTM Engineer job postings are up 205% YoY. But few stacking that title actually have the skills to hold it. I got tired of watching operators call themselves “AI-native” after connecting two Zapier steps. So here's what the real progression looks like, the tech stack, and what’s trending. 𝗚𝗧𝗠 𝗘𝗡𝗚𝗜𝗡𝗘𝗘𝗥 𝗣𝗛𝗔𝗦𝗘𝗦 Phase 1: SDR: Master outreach fundamentals. Learn what gets replies. Phase 2: Operator: Run campaigns end-to-end. Own the full sequence. Phase 3: Systems Builder: Design workflows. Connect tools. Automate repetitive work. Phase 4: Orchestrator: Direct AI agents. Build classifiers. Automate judgment calls. Phase 5: GTM Engineer: Architect the stack. Build systems the existing tools can't handle. A lot of operators I meet are still working through the first three phases. The next skill is true AI-native orchestration to unlock 100x leverage. 𝗧𝗛𝗘 𝗚𝗧𝗠 𝗘𝗡𝗚𝗜𝗡𝗘𝗘𝗥 𝗦𝗧𝗔𝗖𝗞 ↳ ScaledMail – email inboxes ↳ AI Ark – 400M+ contact database ↳ Exa – AI web search ↳ Prospeo – contact enrichment ↳ Hermes – autonomous AI agent ↳ Claude – AI model ↳ HeyReach – LinkedIn sequencer ↳ Attio – AI-native CRM ↳ n8n – automation orchestration ↳ Apollo.io – broad contact database ↳ RevyOps – outbound OS ↳ Clay – GTM workflows ↳ Claude Code – vibe-coding interface ↳ FullEnrich – waterfall enrichment ↳ Obsidian – knowledge base ↳ Slack – comms & automation interface ↳ EmailBison – email sequencer 𝗧𝗥𝗘𝗡𝗗𝗦 𝗜𝗡 𝗚𝗧𝗠 "𝘎𝘛𝘔 𝘌𝘯𝘨𝘪𝘯𝘦𝘦𝘳" 𝘱𝘰𝘴𝘵𝘪𝘯𝘨𝘴 𝘶𝘱 205% → Job listings for the title more than doubled year-over-year, from ~1,400 in mid-2025 to 3,000+ by January 2026. It's one of the fastest-growing titles in B2B tech. 𝘈𝘨𝘦𝘯𝘵𝘴 𝘰𝘷𝘦𝘳 𝘔𝘊𝘗, 𝘯𝘰𝘵 𝘩𝘢𝘯𝘥-𝘴𝘤𝘳𝘪𝘱𝘵𝘴 → The 2026 shift: GTM engineers are delegating work to AI agents via MCP instead of hand-writing every script, orchestration is replacing manual scripting. 𝘚𝘪𝘨𝘯𝘢𝘭 𝘐𝘯𝘵𝘦𝘳𝘱𝘳𝘦𝘵𝘢𝘵𝘪𝘰𝘯 𝘐𝘴 𝘛𝘩𝘦 𝘕𝘦𝘸 𝘓𝘢𝘺𝘦𝘳 → Teams are encoding objection-handling into agents that execute inside guardrails, the Hermes-shaped layer is becoming standard, not novel. 4–5 𝘐𝘯𝘵𝘦𝘨𝘳𝘢𝘵𝘦𝘥 𝘛𝘰𝘰𝘭𝘴 𝘉𝘦𝘢𝘵𝘴 12+ → The 2026 State of GTM Engineering report found lean, tightly integrated stacks consistently outperform teams stitching together a dozen-plus disconnected tools. The "GTM engineer" role is here to stay. What’s going to differentiate top 1% GTM engineers from the rest is the ability to avoid the noise and double-down on the signal for what’s actually making an impact.

The best go-to-market tech stack is the one your reps never have to leave. When you use Clay to build outbound lists, Claude to draft account plans, or LeanData to route hot leads, you want speed. But those tools are only as good as the data powering them. That is why we built ZoomInfo to run natively under the hood of the GTM applications you already use. To make this seamless, we rely on an incredible network of experts. We’re proud to celebrate our amazing cohort of ZoomInfo Solutions Partners: eCore, Iron Horse, Partner UP, Quantum Business Solutions, RevenueHoop, Skaled, SpringDB, and SR Pro. These are the elite agencies, consultants, and GTM architects who help companies wire our verified data directly into their daily workflows. If you are looking to optimize your use of ZoomInfo data, these are the partners to talk to first. Check out all our partners here: https://okt.to/j0TdBA

Clay. Claude. Instantly. Smartlead. Learn these properly and you don't just learn "tools." You end up understanding the entire outbound strategy underneath them. 1- Clay teaches you how enrichment actually works, why one data provider isn't enough, why a waterfall exists, and why "verified" doesn't always mean verified. 2- Claude teaches you how personalization actually works, the difference between a generic AI line and one that proves you actually looked at the company. 3- Instantly and Smartlead teach you how sending infrastructure actually works, deliverability, and warmup and why volume without a ramp burns your domain instead of building it. None of these tools are the strategy. They're what forces you to learn the strategy, because you can't use them well without understanding why each step exists. Most people learn a tool's buttons. Fewer people learn what the tool is actually protecting them from. Heyreach's next on my list to actually dig into!! Anyone here used it enough to tell me what surprised them most? Also genuinely curious from this community: what other tools should I be learning? And if you've been in this space longer than me, am I learning this the right way, or is there something you'd do differently? #GTMEngineering #Clay #Outbound #ColdEmail #B2BSales #SalesTools

Hi all, I haven’t been posting much, but wanted to drop in with an update on my summer project list. Project 1) Cleaned up and repaired a Salesforce instance that had more duplicate customer records than actual customer records (Claude Code, Clay). The dedupe was one piece of a much longer engagement, and most of the work sat upstream of it: understanding how the duplicates were being created, and making sure they stopped. Project 2) First we built a fully custom ICP Engine in Salesforce via Clay (lead/contact/account scores + recurring signsls). Then we rebuilt their GTM lifecycle stages (reo.dev + marketo) and plugged all that data back into Clay audiences for retargeting (Clay Enterprise). Team-up with The Sales Nerd. Project 3) Automated monthly board reporting for a PE-backed client, via Claude Code with a custom HubSpot integration (scoped, read only). We handled their Salesforce > HubSpot migration in Q2, and they found out the hard way that HubSpot reporting and dashboarding tools are clunky at best, and he was previously covering the gap by pulling numbers together by hand. He’s a new dad too, so he no longer has weekend hours to spend building the same deck over and over. Project 4) This was a fun one, also a team-up with The Sales Nerd. We replaced a large data vendor ($200k annual contract) with a custom Clay + Salesforce integration that fully powers their ABM strategy, replicating 100% of prior functionality and then customizing further in directions they previously couldn’t reach: custom signals, fit detection scores, custom matching scenarios. Project 5) An AI science project that grew legs. After reading about the recent progress in mathematics using Claude, I decided to test the limits and use my Claude / Claude Code skills to poke at an unsolved problem, expecting to spend a session or two learning about why it was actually impossible. Instead I ended up digging through decades worth of math research and building an extremely complicated C-based calculator (thanks CC) that now computes a good deal further out than the published tables go. Which, when I look at the list, is the same job as the other four. Every one of these started as data sitting in a form nobody could read, and the work was doing the research to understand it and then building the thing that makes it legible. Four times that was a client’s CRM. The fifth time it was a lattice model from the 1940s. I’m five drafts into turning #5 into a proper paper after a physics professor told me to write it up and submit my research to a journal. More on that soon!

Outbound success relies on a few simple things: 1. A system you can repeat and track 2. A real understanding of your prospect's problem 3. The skills to show your solution instead of just describing it 4. Timing and knowing when a signal makes a message worth sending Everything else is just noise. Tools help but if your stack needs a dozen apps just to run outreach, you'll spend more time managing the stack than actually reaching out and building connections. At Qualeady, we keep the stack small on purpose. Here are the basic tools we actually run on: - Apollo.io to find accounts showing real intent. - Clay for building and enriching prospect lists. - Claude Code for scoring, segmenting, and internal tools. - Smartlead to run email sequences. - OpenPhone as our dialer for calls. - And as a treat, RB2B to see who visits the website, especially the prospects we're targeting. We prefer running LinkedIn outreach manually. It's slower, but less risk of getting flagged or banned, and more control over what actually gets sent to each prospect. This whole stack can be managed by one rep. But if you've got a bigger team and want reps focused purely on outreach, it's worth hiring or assigning a GTM engineer or agency to build automations and orchestrate the list building and enrichment… while the rep runs outreach. I've talked to sales leaders who've been doing it for 20 and 30 years and they all say that the environment has changed a lot but the fundamentals are still the same. Get the system right, understand the problem, show don't tell, and time it well. That's what builds pipeline.

The best go-to-market tech stack is the one your reps never have to leave. When you use Clay to build outbound lists, Claude to draft account plans, or LeanData to route hot leads, you want speed. But those tools are only as good as the data powering them. That is why we built ZoomInfo to run natively under the hood of the GTM applications you already use. To make this seamless, we rely on an incredible network of experts. We’re proud to celebrate our amazing cohort of ZoomInfo Solutions Partners: eCore, Iron Horse, Partner UP, Quantum Business Solutions, RevenueHoop, Skaled, SpringDB, and SR Pro. These are the elite agencies, consultants, and GTM architects who help companies wire our verified data directly into their daily workflows. If you are looking to optimize your use of ZoomInfo data, these are the partners to talk to first. Check out all our partners here: https://okt.to/mGC2Kz

Biblioteca supremă de agenți AI pentru GTM. Într-un singur doc. 50+ resurse alese pe mână. Salveaz-o. 👇 📹 Videos: 20 AI agents replaced our sales team (Jason Lemkin) · Sales "System of Action" (Clay, Sequoia) · AI SDRs, 6 months later (SaaStr) +6 📁 Repos: Anthropic Agent Skills (170.5k⭐) · GTM Skills (322 skills open-source) · gtm-agents pt Claude Code (378⭐) +5 📖 Guides: Building Effective Agents (Anthropic) · Practical Guide to Agents (OpenAI) · GTM Engineering (Clay) +6 📚 Books: Revenue Architecture · Founding Sales · Obviously Awesome +3 📊 Reports: 2025 GTM Benchmarks (Pavilion x Ebsta) · AI Won't Transform Your GTM (Bain) · Anthropic Economic Index +4 🎓 Courses: Clay University · GTM Engineer School · Agentic AI — Andrew Ng (DeepLearning.AI) +7 Comentează "BIBLIOTECA" și îți trimit tot doc-ul în DM. ♻️ Repost dacă cineva din rețeaua ta are nevoie.

poe.com/JobsExpert #chatbot #ai #anthropic #gpt #openai #mistral #claude #perplexity #chat #virtual #agentic #agent #LLM #machinelearning #aitrainer #conversationalai #chatbotpersona #conversationdesign #meds #medicaid #zorg #medical #livesupport #breathe #medicalresearch #doctor #surgeon #surgery Sam Altman Dario Amodei #fashion #robot #makeup #mask #glitter #jewels #gold #couture #style #stylish #catwalk #model #runway #glamour #makeup #pearls #tattoos #diamonds #cristal #luxurious #glam #elite #rich #samaltman #amodei #darioamodei #DataAnnotation #TechCareers Anthropic #OpenAI #TrustAndSafetyProfessionals #LinkedInSEO #LinkedInOptimization #LinkedInGrowth #LinkedInBranding #LinkedInNetworking #LinkedInVisibility #LinkedInProfessional #LinkedInEngagement #LinkedInContent #elonmusk #jeffbezos #billgates #sundarpichai #markzuckerberg #timcook #satyanadella #donaldtrump #barackobama #joebiden #warrenbuffett #larrypage #sergeybrin #peterthiel #reidhoffman #marissamayer #sherylsandberg #richardbranson #oprahwinfrey #jackdorsey #jensenhuang #lisasu #RuleTheWorld #WorldOracle #Leadership #ExecutiveWisdom #DecisionMaking #GlobalStrategy #ThoughtLeadership #Influence #Visionary #HumanInsights #DailyGuidance #StrategicAdvice #michaelbloomberg #georgehotz #navalravikant #chamathpalihapitiya #vitalikbuterin #czbinance #andrewyang #yanlecun #geoffreyhinton #demis #karpathy #lexfridman #emadmostaque #jimfan #karim #benhorowitz #marcandreessen #paulkagame #katiehaun #balajis #garrytan #patrickcollison #johncollison #danielgross #emilychang #caseyneistat #kevinrose #kevinhart #taylorswift #kanyewest #drake #rihanna #beyonce #ladygaga #madonna #lebronjames #michaeljordan #serenawilliams #lionelmessi #cristianoronaldo #neymar #rogerfederer #tigerwoods #tombrady #stephcurry #viratkohli #elonmuskAI #bezosAI #trumpAI #gatesAI #nadellaAI #zuckerbergAI #altmanAI #huangAI #buterinAI #andreessenAI #Anthropic #OpenAI #Google #Microsoft #Meta #xAI #Cohere #MistralAI #Perplexity #DeepMind #StabilityAI #HuggingFace #Midjourney #Runway #CharacterAI #Poe #Quora #iMerit #Welocalize #Appen #NVIDIA #IBM #Oracle #Snowflake #CoreWeave #Databricks #Palantir #ScaleAI #ElevenLabs #Glean #Harvey #HeyGen #Cognition #WorldLabs #Gamma #ChaiDiscovery #Rogo #Abridge #AppliedIntuition #Baseten #BlackForestLabs #Clay #Crusoe #Cursor #Cyera #Decagon #EliseAI #Fal #FireworksAI #Genspark #Krea #Legora #ListenLabs #Lovable #Mercor #OpenEvidence #PhysicalIntelligence #ReflectionAI #SafeSuperintelligence #Anduril #Cerebras #Groq #InflectionAI #Waymo #Writer #Writesonic #Zoox #AlephAlpha #Baidu #SenseTime #Adobe #Salesforce #AmazonAI #AMD #Intel #Qualcomm #C3ai #PalantirTechnologies #UiPath #DataRobot #H2Oai #TensorFlow #PyTorch #LangChain #LlamaIndex #Pinecone #Weaviate #Milvus #Qdrant #ChromaDB #DeepL #Synthesia #ElevenLabs #Descript #AssemblyAI #Deepgram #Voiceflow #Botpress

https://lnkd.in/g6HNG92M Evidence Before Conclusions: A Challenge to Tom Brown, Anthropic Co-Founder. Continuity involving OpenAI GPT‑2, GPT‑3, scaling, safety, interpretability, infrastructure, and human-feedback research. The coordinated seven-person departure. Extensive movement of OpenAI personnel. Anthropic’s early dependence on knowledge and experience accumulated by former OpenAI personnel. Chronology conflicts and unanswered questions surround early development. The world deserves transparency about Claude’s AI provenance. This question belongs to policymakers, lawmakers, investors, universities, healthcare leaders, Bob Sternfels, Jamie Dimon, Mark Hussey, Greg Brockman, Scott Pulsipher, and Gene Hayes. Can Anthropic separate Claude’s technical and organizational DNA from OpenAI, GPT‑2, GPT‑3, and its former OpenAI personnel? Evidence Anthropic research: https://lnkd.in/gY8vy_u4 Full paper: https://lnkd.in/g2gWvNKd OpenAI’s 2017 human-preferences research: https://lnkd.in/gU2aj2Xf OpenAI Frontier Alliances: https://lnkd.in/g456aV4Z Independent provenance record: https://lnkd.in/gRazuq4T Would Anthropic—or Claude in its present form—exist without the founders’ experience at OpenAI, scaling and safety research, infrastructure knowledge, and the former OpenAI employees Tom Brown described? What came from public OpenAI science? What was portable professional knowledge? What did Anthropic independently develop? Can dated records prove those boundaries? Anthropic reports progress across ten alignment-failure categories—and cheating behavior in 39 of 1,601 research trajectories. Humans still choose the benchmarks, rules, evaluators, and definition of success. Similarity does not prove copying, and ancestry does not prove unlawful use. But leadership claims cannot replace independent, object-specific provenance review. Evidence matters. Transparency matters. Truth must be tested. Interesting! Jeff Winter, Amir Hartman, Gregory Charlop, MD, DipABLM, Clay Magouyrk, Ammar S., Dr. Osama Salman Mohammed, Ashley Nicholson, Ralph Aboujaoude Diaz, Stefano M. Sinicropi, MD, Dr. Saleh ASHRM - iMBA Mini, Holland Haynie, MD, Dr Nik, James Kaplan, Eric Kimberling

Running cold email has become expensive and harder in last year.. Here's what you get with Lead Bakery: 🔹 Instantly account, fully set up with claude MCP 🔹 Unlimited leads for your niche via GetLeads.io 🔹 Personalization via Clay 🔹 50-100 Google Workspace inboxes 🔹 2k/day sending volume All done for you, A to Z — for under $200/month + per lead. No more going down or just not knowing what to do. Just a cold email system that runs itself. we can work with any B2B company.. as long as you have a large enough TAM. you just log in and watch the calls hit your calendar. 👉 Check us out: leadbakery.io

If you're a B2B founder paying $25K/year for an AI SDR, read this. Mistake we made: We bought volume without judgment. AI SDR tool blasted 2,000 emails/week. Reply rate dropped from 8% to 1.2%. Domain reputation tanked. Lesson: The 3 AI traps killing B2B marketing in 2026: Trap 1: AI SDR Trap — volume without judgment Trap 2: Content Slop Trap — 91% of teams increased output, only 39% saw better performance Trap 3: No Governance Trap — generic AI content gets 30% less reach on LinkedIn Fix that worked: → Clay enriches + scores leads (replaces 5 tools) → Claude Code writes personalized first line based on 10-K / LinkedIn activity → Human AE reviews top 20% only Result: 47% more productive sales pros, saving 12 hours/week, with 83% of AI-enabled teams seeing revenue growth. Don't automate bad outreach. Fix the workflow, then automate. Want my Clay scoring prompt? Comment CLAY.

What would actually break if you cancelled every outbound tool you pay for? I found out last quarter. $1,107 a month — Clay, Phantombuster, Apollo, HeyReach, Sales Nav — turned off one at a time over two months. Nothing important broke. Reply rates went up. The stack wasn't solving 2026's problem. It was solving 2018's, expensively — and the two things that replaced most of it were Claude and a spreadsheet, not another ChatGPT wrapper. Here are the 6 Claude prompts that replaced it 👇 [Save for later] 1. Phantombuster, $50 It scraped 5,000 profiles. So does every other rep, and they're the same 5,000. The list stopped being the asset — the person who moved this week is. Paste into Claude: "Act as a researcher. From this list, return only the accounts where something changed in the last 7 days. Discard the rest. [paste list]." 2. Clay, $800 Forty enrichment columns nobody read, replaced by one 0-100 score with a hard gate at 80. The gate is the whole thing. Paste into Claude: "Act as a scoring analyst. Reduce these 40 columns to the 4 that actually predict a reply, and score fit plus urgency 0-100. [paste data]." 3. Apollo, $99 A contact database, for the 20% who don't reply on LinkedIn. Paste into Claude: "Act as a research assistant. For these non-responders, list the public places a work email would appear. Don't guess an address. [paste names]." 4. HeyReach, $79 Five-step sequences, blasted. Same leads: 4% reply through the sequence, 47% through one hand-written signal DM. Paste into Claude: "Act as an outbound writer. Take this 5-step sequence and replace it with one message built on this signal. Tell me what the other four steps were compensating for. [paste sequence]." 5. Sales Nav, $79 Boolean searches, replaced by a signal calendar — watch the 6 of your 200 who did something this week. Paste into Claude: "Act as my SDR. From these 200 accounts, name the handful with a public event this week and put them in date order. [paste accounts]." 6. The 2018 Test The question I now run on every line item, and it's uncomfortable on about half of them. Paste into Claude: "Act as a skeptical CFO. For each tool, what problem is it solving in 2026? If the honest answer is the 2018 version, say cancel. [paste stack]." Old stack: $1,107/mo, about 2% reply. New loop: a spreadsheet, Claude and LinkedIn, 40-60% on scored signals. What's the one tool you keep paying for purely out of habit? ----- Full breakdown, free and ungated — link in the first comment. #Outbound #B2BSales #Claude #LinkedInStrategy

Marketing in 2026 isn’t about using more tools. It’s about using the right tools—at the right time. 🚀 The marketing landscape has changed dramatically. Earlier, we relied heavily on: 🔹 Google Search for research 🔹 Gmail for communication 🔹 Google Ads & Meta Ads for acquisition 🔹 WordPress for websites 🔹 Excel/Sheets for reporting Today, AI and automation have completely changed the workflow. From ChatGPT, Claude & Gemini for research and strategy to Zapier & Make for automation, and Figma, Cursor, Notion, Loom & Clay for faster execution. But one thing hasn’t changed: 👉 Google Ads still matters. 👉 Meta Ads still matters. 👉 Google Analytics still matters. The tools are evolving, but marketing fundamentals remain the same. Understanding your audience. Creating the right offer. Tracking the numbers. Optimizing based on data. And ultimately, generating profitable results. The future belongs to marketers who can combine AI + data + strategy. What marketing tool has changed your workflow the most in 2026? 👇 #Marketing #DigitalMarketing #AI #PerformanceMarketing #Marketing2026 #GoogleAds #MetaAds #AITools #GrowthMarketing #RaviYadav

Hypergen - B2B Cold-Email is hiring a GTM Engineer who talks to Claude more than to their coworkers. Hypergen builds outbound for 120+ B2B clients. Cold email, Clay, RevOps. Where: remote, anywhere in Europe, South Africa, or Canada. What you'll own: - Claude and Clay automations and enrichment workflows - Claude skills and agents that research, write, and QA at scale - Outbound campaigns end-to-end What we need: - Hands-on Clay experience (non-negotiable) - Real Claude usage, I will ask you to open up your GitHub repo or show us examples during the interview - Strong B2B cold email background What you get: - Competitive salary + performance bonuses - Budget for courses and training, on us - Remote-first, flexible hours - A team that argues about ICPs during lunch DM me or send your CV to alex@hypergen.io Subscribe: https://lnkd.in/dBZvz9Bc When reaching out, mention that you saw this message via GTM Daily.

If I were building a GTM engineering system in 2026, these are the tools I'd connect: - HubSpot / Salesforce → closed-won + closed-lost data to understand what actually converts - Clay + ChatGPT + Claude → analyse customer data and build a more accurate ICP - Clay + LinkedIn Sales Navigator + Ocean.io → map the TAM instead of relying on one database - Perplexity + LinkedIn + company websites → research accounts and add useful context - Clay + LinkedIn Sales Navigator → enrich accounts and identify the right people - HubSpot / Salesforce / Attio → keep the CRM as the system of record - Clay + AI → qualify, score and route accounts based on ICP + situation + signals - Awareness scoring → understand whether an account is unaware, aware, researching, considering or buying - Outbound + inbound + ABM → activate the right GTM play for the account and situation - HubSpot / Salesforce → connect meetings and sales activity back to closed-won revenue Every tool has a specific role. And every layer should: - Improve the data - Improve the targeting - Improve the context - Improve the next decision (and no, adding 15 more tools doesn't magically create a GTM system 😂) The goal isn't to have the biggest stack. It's to make the stack work together: customer data → ICP → TAM → research → qualification → signals → GTM play → revenue What would you add to this system?

Funnel IA complet - relier les outils à un objectif business. TOFU - Attirer l'attention Outils : Claude/ChatGPT · Canva · OpusClip · Runway · Buffer Actions : scripts · visuels · clips courts · publication MOFU - Capturer et qualifier Outils : Tally · Typeform · ManyChat · Beehiiv · Chatbase · HubSpot Actions : formulaire · DM auto · newsletter · qualification BOFU - Convertir Outils : Clay · Apollo · Lemlist · Calendly · Stripe · Make Actions : prospection enrichie · séquences · RDV · paiement Delivery - Livrer Outils : Notion · Softr · Glide · Airtable · Chatbase · Slack Actions : documentation · portail client · support · suivi Exemple complet : 1. Claude génère 5 posts LinkedIn depuis un cas client 2. Les commentateurs reçoivent un guide via ManyChat 3. Tally collecte leur contexte 4. Airtable stocke et score les leads 5. Claude génère un résumé personnalisé 6. Calendly propose un RDV aux leads chauds 7. Stripe encaisse un audit payant 8. Notion livre la synthèse 9. Chatbase répond aux questions fréquentes La compétence clé n'est pas de tout déléguer. C'est de savoir cadrer, contrôler et intégrer l'IA. Commente FUNNEL IA → mon architecture complète #FunnelIA #TOFU #MOFU #BOFU #ManyChat #Claude #Stripe #IA2026 #NextGenDevHub #DéveloppeurIA #Acquisition

They tell us Claude is #unbiased, yet the company's inner #circle is deeply #intertwined with #Silicon Valley's #elite power #brokers. Sam Altman Dario Amodei #fashion #robot #makeup #mask #glitter #jewels #gold #couture #style #stylish #catwalk #model #runway #glamour #makeup #pearls #tattoos #diamonds #cristal #luxurious #glam #elite #rich #samaltman #amodei #darioamodei #DataAnnotation #TechCareers Anthropic OpenAI #TrustAndSafetyProfessionals #LinkedInSEO #LinkedInOptimization #LinkedInGrowth #LinkedInBranding #LinkedInNetworking #LinkedInVisibility #LinkedInProfessional #LinkedInEngagement #LinkedInContent #elonmusk #jeffbezos #billgates #sundarpichai #markzuckerberg #timcook #satyanadella #donaldtrump #barackobama #joebiden #warrenbuffett #larrypage #sergeybrin #peterthiel #reidhoffman #marissamayer #sherylsandberg #richardbranson #oprahwinfrey #jackdorsey #jensenhuang #lisasu #RuleTheWorld #WorldOracle #Leadership #ExecutiveWisdom #DecisionMaking #GlobalStrategy #ThoughtLeadership #Influence #Visionary #HumanInsights #DailyGuidance #StrategicAdvice #michaelbloomberg #georgehotz #navalravikant #chamathpalihapitiya #vitalikbuterin #czbinance #andrewyang #yanlecun #geoffreyhinton #demis #karpathy #lexfridman #emadmostaque #jimfan #karim #benhorowitz #marcandreessen #paulkagame #katiehaun #balajis #garrytan #patrickcollison #johncollison #danielgross #emilychang #caseyneistat #kevinrose #kevinhart #taylorswift #kanyewest #drake #rihanna #beyonce #ladygaga #madonna #lebronjames #michaeljordan #serenawilliams #lionelmessi #cristianoronaldo #neymar #rogerfederer #tigerwoods #tombrady #stephcurry #viratkohli #elonmuskAI #bezosAI #trumpAI #gatesAI #nadellaAI #zuckerbergAI #altmanAI #huangAI #buterinAI #andreessenAI #Anthropic #OpenAI #Google #Microsoft #Meta #xAI #Cohere #MistralAI #Perplexity #DeepMind #StabilityAI #HuggingFace #Midjourney #Runway #CharacterAI #Poe #Quora #iMerit #Welocalize #Appen #NVIDIA #IBM #Oracle #Snowflake #CoreWeave #Databricks #Palantir #ScaleAI #ElevenLabs #Glean #Harvey #HeyGen #Cognition #WorldLabs #Gamma #ChaiDiscovery #Rogo #Abridge #AppliedIntuition #Baseten #BlackForestLabs #Clay #Crusoe #Cursor #Cyera #Decagon #EliseAI #Fal #FireworksAI #Genspark #Krea #Legora #ListenLabs #Lovable #Mercor #OpenEvidence #PhysicalIntelligence #ReflectionAI #SafeSuperintelligence #Anduril #Cerebras #Groq #InflectionAI #Waymo #Writer #Writesonic #Zoox #AlephAlpha #Baidu #SenseTime #Adobe #Salesforce #AmazonAI #AMD #Intel #Qualcomm #C3ai #PalantirTechnologies #UiPath #DataRobot #H2Oai #TensorFlow #PyTorch #LangChain #LlamaIndex #Pinecone #Weaviate #Milvus #Qdrant #ChromaDB #DeepL #Synthesia #ElevenLabs #Descript #AssemblyAI #Deepgram #Voiceflow #Botpress #ManyChat #LivePerson #JasperAI #Copyai #WriterAI #Grammarly #NotionAI #Canva #Phind #PerplexityAI #Grok #Duckai #FreedomGPT #Ollama #LMStudio #VeniceAI #ElevenLabs

J'ai fini par rassembler la liste ultime. Pas celle basée sur la hype, ni sur ce qui fait le buzz sur X, mais celle des outils que j'utilise vraiment dans mes workflows, tous les jours. J'ai tout structuré en 6 catégories pour que vous trouviez exactement ce qui correspond à votre cas d'usage : - Assistants Généraux : Claude, ChatGPT, Perplexity - Développement : Cursor, Lovable, Replit, Bolt - Création de Contenu : Manus AI, HeyGen, Synthesia, Descript, Opus Clip, Beehiiv - Productivité : Grammarly, NotebookLM, Otio AI, Gamma, Granola, Superhuman, Wispr Flow - Créativité : ElevenLabs, Suno, Midjourney, Runway, Kling, Pika Labs, Figma, Canva, Google Veo, Higgsfield - Automatisation : Softr, n8n, Zapier, Lindy AI, Claude Code, Chatbase, Gemini, Notion AI, Apify, Clay Quelques leçons tirées après en avoir testé des centaines : Les outils qui restent sont ceux qui s'intègrent à ce que vous faites déjà. Si ça prend 2 heures à configurer et que vous ne l'ouvrez plus jamais, sa puissance n'a aucune importance. Le plus gros changement que j'observe en 2026, c'est que les outils de code sont devenus accessibles à tous. Cursor, Lovable et Replit permettent aux fondateurs de créer et lancer des produits sans écrire de code en partant de zéro. C'était impensable il y a encore 18 mois. Et côté vidéo, Google Veo est discrètement devenu le meilleur modèle de génération vidéo IA du marché. La plupart des gens parlent encore des anciens modèles pendant que Veo a pris une avance monumentale. Je continuerai de mettre à jour cette stack au fil des sorties. À vous : quelle catégorie d'outils IA vous a fait gagner le plus de temps cette année ? 🔖 Sauvegardez ce post pour plus tard ♻️ Pensez à reposter pour en faire profiter votre réseau. ___ PS : Vous souhaitez intégrer l'IA rapidement dans votre entreprise ? Je propose des créneaux d'échange de 30 minutes pour en discuter : → https://lnkd.in/dSsrWQe2

Most founders chase $1M ARR with Random tools. 𝗪𝗿𝗼𝗻𝗴 𝗺𝗼𝘃𝗲. The founders who actually hit that number don't use more tools. They use the RIGHT seven. (📌 𝗦𝗮𝘃𝗲 𝘁𝗵𝗶𝘀 before you build your next system.) ➡️Here's the exact stack that gets you there:👇 ☑️𝟭. 𝗟𝗢𝗩𝗔𝗕𝗟𝗘 Stop waiting on your dev team for every test. - Design a prototype in hours, not weeks - Launch and validate before you commit budget - Kill bad ideas early, cheap This is how you move at founder speed. ☑️𝟮. 𝗦𝗘𝗔𝗥𝗖𝗛𝗔𝗕𝗟𝗘 SEO used to mean guessing and waiting months. Now it runs on autopilot: - Tracks SEO, AEO, and GEO together - Adjusts before rankings even drop - Frees you from checking dashboards daily Your growth assistant that never clocks out. ☑️𝟯. 𝗖𝗨𝗥𝗦𝗢𝗥 Your engineering team is your biggest cost center. Cursor cuts that cost: - Writes code alongside your devs, not for them - Cuts review and debug time hard - Ships features faster without new hires Speed here compounds everywhere else. ☑️𝟰. 𝗭𝗔𝗣𝗜𝗘𝗥 Manual customer follow-up doesn't scale past 100 users. - Connects your CRM to every other tool - Automates retention sequences - Triggers upsell campaigns without a human touch This is the difference between managing customers and losing them. ☑️𝟱. 𝗖𝗛𝗔𝗧𝗚𝗣𝗧 + 𝗖𝗟𝗔𝗨𝗗𝗘 Generic onboarding loses new hires fast. - Train each one on your actual team voice - Build support docs that sound like you - Personalize training without hiring a trainer Two tools. One consistent brand voice at scale. ☑️𝟲. 𝗖𝗟𝗔𝗬 Sales teams waste hours enriching contact lists by hand. - Pulls social, sales, and marketing data into one view - Enriches leads automatically at scale - Hands your team warm context, not cold names Better data means fewer wasted calls. ☑️𝟳. 𝗖𝗟𝗔𝗨𝗗𝗘 𝗖𝗢𝗗𝗘 Churn kills ARR faster than any competitor does. - Flags churn signals before they show in your reports - Analyzes trends across your entire customer base - Automates the follow-up communication Catching churn early is cheaper than replacing revenue. Here's the pattern: ↳ Lovable → build fast ↳ Searchable → grow visibility ↳ Cursor → ship code ↳ Zapier → automate ops ↳ ChatGPT/Claude → train your team ↳ Clay → enrich your leads ↳ Claude Code → cut your churn Seven tools. Seven bottlenecks removed. Most founders think $1M ARR needs more effort. It needs fewer gaps in the system. 𝗥𝗲𝗽𝗼𝘀𝘁♻️ this if it helped ps: Which one of these are you not using yet? Comment below. ____ Nikhil

El trabajo de growth en 2026 no se parece en nada al de hace dos años. Antes todo corría sobre una lista corta de herramientas. SEO era Google. Paid era Meta y Google. Los datos salían de ZoomInfo. El sitio estaba en WordPress. Los reportes eran Google Analytics 4 y una planilla. Las automatizaciones, si había, eran Zapier. El outbound era mandar un mail desde Gmail. Ahora mirá sobre qué corre el mismo trabajo. Ya no te buscan solo en Google: también en ChatGPT, Claude y Gemini. Los reportes viven en PostHog y HockeyStack. El outbound corre en Instantly.ai, lemlist y HeyReach.io La automatización, en Clay, n8n y Make. El paid se expandió a LinkedIn, Reddit, Inc. y X. El sitio se construye con AI en Framer o Vercel. Pero el cambio real es otro. Es cuánto puede hacer una sola persona. Una idea del lunes a la mañana puede estar viva el mismo día. Investigar el mercado. Armar la lista de contactos. Publicar una landing. Hacer los creativos y lanzarlos. Y ver los primeros números a la tarde, sin esperar a nadie. Ese es el cambio grande, porque el trabajo viejo estaba lleno de dependencias. Entender un mercado nuevo eran días de research. Una landing nueva era esperar a un desarrollador. Datos limpios eran horas de trabajo manual. El stack nuevo se llevó casi todo eso. Y por eso cambia lo que te hace bueno en este trabajo. Ser excelente en un canal sigue contando. Pero hoy todos tienen las mismas herramientas, así que ahí la ventaja dura poco. Lo que separa es armar el sistema completo y ponerlo en el aire antes de que se pase el momento. Antes ganabas por saber más que el otro. Ahora ganás por llegar antes. --- 👉 P.S. Armar el sistema completo, y no solo dominar un canal, es lo que ordena el Curso Growth Rockstar: modelos de growth, adquisición, retención y monetización aplicados a tu negocio. Por ahí ya pasó gente de Rappi, Mercado Libre, Nubank y Tiendanube. La edición 15 arranca el 14 de septiembre y quedan los últimos cupos para aplicar. Comentá "GROWTH" y te mando el link

Probé decenas de herramientas de IA este año. Estas 10 son las que quedaron. Una por tarea, sin repetir función: 1 - ChatGPT. El todoterreno del día a día. 40 variantes de anuncio en 10 minutos, en vez de 4 en una tarde. 2 - Claude. Texto largo y análisis a fondo. 30 entrevistas a clientes convertidas en un mapa de objeciones con citas textuales. 3 - Gemini. Datos dentro de Google. Limpia, segmenta y puntúa tu base de leads sin sacarla de la hoja. 4 - Perplexity. Investigación con fuente al lado. Competencia, precios y quién decide, verificable en un clic. 5 - Clay. Prospección enriquecida. De 1.000 contactos genéricos a 80 con un motivo real para escribirles. 6 - Apollo.io. Base de contactos B2B. Correo, cargo y empresa de tu lista objetivo en minutos. 7 - n8n. Automatización de flujos. Conecta formulario, CRM y Slack sin escribir una línea de código. 8 - Gamma. Presentaciones y propuestas. Una propuesta de cliente lista en 15 minutos, no en dos días. 9 - Canva. Piezas gráficas. 20 creativos para tu próximo test de anuncios sin pasar por diseño. 10 - Fathom.ai. Llamadas y voz del cliente. Graba cada demo y te deja por escrito las objeciones que salieron. El error común es abrir las 10 al tiempo. Empieza por una: la que resuelva la tarea que más horas te come esta semana. Compartimos este tipo de contenido cada semana en La Liga del Growth, una comunidad de WhatsApp de operadores de growth en Latam: https://lnkd.in/eVeUUkNa

The AI winners of 2026 won’t use every app, they’ll build the right stack. Here’s the shortlist. 1 → General assistants → Claude pressure-tests strategy and long-form decisions. → ChatGPT analyzes files and ships routine work. → Perplexity researches markets with cited web answers. 2 → Development and building → Cursor edits entire codebases through prompts. → Replit prototypes and deploys products in one place. → Base44 builds internal tools without infrastructure overhead. 3 → Content creation → Beehiiv publishes and monetizes a founder newsletter. → HeyGen creates presenter videos without a crew. → Opus Clip repurposes long videos into social clips. → Manus AI handles multi-step research and execution. 4 → Productivity → Granola captures meetings without manual notes. → Gamma turns rough ideas into polished decks. → Superhuman triages email with AI-assisted workflows. → Grammarly tightens writing before customers see it. → Otio AI organizes and summarizes research projects. → Wispr Flow turns dictation into drafts faster than typing. 5 → Creativity → Suno generates original music for campaigns. → Canva creates branded visuals without extra headcount. → ElevenLabs produces natural voiceovers and localized audio. → Runway generates and edits campaign video. → Kling creates cinematic video from text or images. → Pika Labs animates ideas into shareable videos. → Figma moves designs into prototypes faster. → Google Veo produces polished scenes from prompts. → Higgsfield creates camera-controlled social video. 6 → Automation and integration → n8n connects tools through flexible workflows. → Clay enriches leads and personalizes outbound. → Claude Code delegates coding inside your terminal. → Softr turns operational data into client portals. → Gemini works across Google files and research. Which 5–6 tools would earn a permanent place in your AI stack? 𝗙𝗥𝗘𝗘 (𝗚𝗼𝗼𝗴𝗹𝗲) 𝗖𝗼𝘂𝗿𝘀𝗲𝘀 𝘆𝗼𝘂 𝘄𝗶𝗹𝗹 𝗿𝗲𝗴𝗿𝗲𝘁 𝗻𝗼𝘁 𝘁𝗮𝗸𝗶𝗻𝗴 𝗶𝗻 𝟮𝟬𝟮𝟲. 1. Google Data Analytics: https://lnkd.in/gmpYBGQz 2. Google Project Management: https://lnkd.in/g2SFqfdw 3. Foundations of Project Management: https://lnkd.in/gAphbnV7 4. Google Introduction to Generative AI: https://lnkd.in/gHnK6GAA 5. Google Cybersecurity: https://lnkd.in/gwT8n2XD 6. Google UX Design: https://lnkd.in/gtifc_sH 7. Google Digital Marketing & E-commerce:https://lnkd.in/gNUkiJN6 8. Google IT Support:https://lnkd.in/gHrnNNrc 9. Web Applications for Everybody:https://lnkd.in/gm5jjAEu 10. Get Started with Python: https://lnkd.in/gE_8qkVJ 11. Learn Python Basics for Data Analysis:https://lnkd.in/gshCUpGM 12. Create your own Python objects:https://lnkd.in/d_rR29MN 13. Data Analysis with R Programming:https://lnkd.in/gDxWYtnD 14. IBM Full Stack Software Developer: https://lnkd.in/gyVWhYXv 15. Introduction to Web Development (HTML, CSS, JS): https://lnkd.in/giSVuNjj 16. IBM Front-End Developer:https://lnkd.in/gBUVNYZv

Anfänger nutzen eine KI als vermeintlichen Allrounder. Experten nutzen nur das beste Tool für die jeweilige Aufgabe. Hört auf, nach der einen KI zu suchen, die gibt's nicht! 🧐 So sieht mein persönlicher Tool-Stack aktuell aus: 🏆 𝐀𝐥𝐥𝐠𝐞𝐦𝐞𝐢𝐧𝐞 𝐀𝐬𝐬𝐢𝐬𝐭𝐞𝐧𝐭𝐞𝐧 Claude, ChatGPT & Gemini 🎥 𝐕𝐢𝐝𝐞𝐨- & 𝐀𝐮𝐝𝐢𝐨-𝐏𝐫𝐨𝐝𝐮𝐤𝐭𝐢𝐨𝐧 Manus AI, HeyGen, Descript & Opus Clip 💻 𝐂𝐨𝐝𝐞𝐛𝐚𝐬𝐞 & 𝐒𝐨𝐟𝐭𝐰𝐚𝐫𝐞𝐞𝐧𝐭𝐰𝐢𝐜𝐤𝐥𝐮𝐧𝐠 Cursor, Lovable, Replit, Base44, Claude Code & GitHub Copilot 🚀 𝐖𝐞𝐫𝐤𝐩𝐥𝐚𝐭𝐳-𝐄𝐟𝐟𝐞𝐤𝐭𝐢𝐯𝐢𝐭ä𝐭 Grammarly, NotebookLM, Otio AI, Granola, Superhuman, Wispr Flow und Gamma 🎨 𝐃𝐞𝐬𝐢𝐠𝐧, 𝐀𝐫𝐭 & 𝐀𝐮𝐝𝐢𝐨𝐞𝐫𝐳𝐞𝐮𝐠𝐮𝐧𝐠 ElevenLabs, Suno, Midjourney, Runway, Kling, Pika Labs, Figma, Canva, Google Veo & Higgsfield ⚙️ 𝐏𝐫𝐨𝐳𝐞𝐬𝐬𝐚𝐮𝐭𝐨𝐦𝐚𝐭𝐢𝐨𝐧 & 𝐀𝐩𝐩-𝐕𝐞𝐫𝐛𝐢𝐧𝐝𝐮𝐧𝐠𝐞𝐧 Softr, n8n, Zapier, Lindy AI, Claude Cowork, Chatbase, ManyChat, Notion AI, Apify & Clay Der entscheidende Unterschied liegt nicht darin, wer die "beste" KI entdeckt hat. Der Unterschied ist, wer aufgehört hat, nach EINER Universallösung zu suchen, und stattdessen für jede Situation das passende Werkzeug einsetzt. Welche Tools sind in eurer Toolkit fest verankert? Und gibt es spezialisierte Lösungen, die ich noch übersehen habe?

Excited to share that Anthropic just highlighted Artemis Security in their Claude Code Guide for Startups, a writeup on how the most AI-native companies operate. It's a privilege to be recognized alongside fellow AI-native companies including Cognition, Clay, and ClickHouse and flattering to be the only cyber security company mentioned. 🚀 🚀 https://lnkd.in/gq8XCUGt

The modern GTM stack is not Clay vs n8n vs Claude Code. The more useful setup is giving each one a job. Clay is still good when I want to see the data, inspect the list and iterate quickly. Automation tools connect steps. Code agents can handle custom scrapers, API workflows and enrichment waterfalls. One operator in the 2026 material described building a scraper with Claude Code, exposing it through an API, then plugging that API back into Clay. That is the direction I find more interesting. Not "which tool wins?" How do the tools work together?

The modern GTM stack is not Clay vs n8n vs Claude Code. The more useful setup is giving each one a job. Clay is still good when I want to see the data, inspect the list and iterate quickly. Automation tools connect steps. Code agents can handle custom scrapers, API workflows and enrichment waterfalls. One operator in the 2026 material described building a scraper with Claude Code, exposing it through an API, then plugging that API back into Clay. That is the direction I find more interesting. Not "which tool wins?" How do the tools work together?

Campaign / Strategist Manager | Lead Gen Jay 📍 Medellín, Colombia | Poblado Office Lead Gen Jay is hiring. 🚀 We’re looking for a Campaign / Strategist Manager to join our growing team in Medellín and take ownership of cold email campaigns and client results. This isn’t a role where you simply launch campaigns and check boxes. We want someone who can think strategically, solve problems, communicate confidently with clients, and use AI + automation to get better results faster. What You’ll Do • Own and manage multiple client campaigns from strategy through execution • Optimize for deliverability, replies, meetings booked, and ROI • Build targeted prospect lists and develop compelling offers • Write and test high-converting cold email copy • Analyze performance and continuously improve campaigns • Communicate proactively with clients and keep them informed • Troubleshoot problems and find solutions quickly • Use AI and automation to improve speed, efficiency, and results What We’re Looking For • 2+ years of experience in campaign management, strategy, cold email, lead generation, or a similar role • Strong understanding of cold email, deliverability, targeting, copywriting, and campaign optimization • Excellent communication and organizational skills • Proactive, resourceful, and comfortable taking ownership • Strong attention to detail • Comfortable managing multiple clients at once • Excited about AI, automation, and finding better ways to work • Able to thrive in a fast-paced, performance-driven environment • Based in Medellín and able to work from our Poblado office Experience with Instantly, Clay, Claude Code, ChatGPT, Apollo, GoHighLevel, n8n, Zapier, or Make is a major plus. Who Will Thrive Here? Someone who doesn't wait around to be told what to do. If something isn't working, you investigate it. If a client needs something, you communicate. If there's a better way to do something, you find it. We’re looking for someone who is smart, proactive, accountable, curious, and obsessed with getting better results. Ready to Apply? --> https://lnkd.in/dmbu7Mnq

Campaign / Strategist Manager | Lead Gen Jay 📍 Medellín, Colombia | Poblado Office Lead Gen Jay is hiring. 🚀 We’re looking for a Campaign / Strategist Manager to join our growing team in Medellín and take ownership of cold email campaigns and client results. This isn’t a role where you simply launch campaigns and check boxes. We want someone who can think strategically, solve problems, communicate confidently with clients, and use AI + automation to get better results faster. What You’ll Do • Own and manage multiple client campaigns from strategy through execution • Optimize for deliverability, replies, meetings booked, and ROI • Build targeted prospect lists and develop compelling offers • Write and test high-converting cold email copy • Analyze performance and continuously improve campaigns • Communicate proactively with clients and keep them informed • Troubleshoot problems and find solutions quickly • Use AI and automation to improve speed, efficiency, and results What We’re Looking For • 2+ years of experience in campaign management, strategy, cold email, lead generation, or a similar role • Strong understanding of cold email, deliverability, targeting, copywriting, and campaign optimization • Excellent communication and organizational skills • Proactive, resourceful, and comfortable taking ownership • Strong attention to detail • Comfortable managing multiple clients at once • Excited about AI, automation, and finding better ways to work • Able to thrive in a fast-paced, performance-driven environment • Based in Medellín and able to work from our Poblado office Experience with Instantly, Clay, Claude Code, ChatGPT, Apollo, GoHighLevel, n8n, Zapier, or Make is a major plus. Who Will Thrive Here? Someone who doesn't wait around to be told what to do. If something isn't working, you investigate it. If a client needs something, you communicate. If there's a better way to do something, you find it. We’re looking for someone who is smart, proactive, accountable, curious, and obsessed with getting better results. Ready to Apply? --> https://lnkd.in/dRHxVXwg

Just got off a session with Kushagra T. on GTM Engineering and this was honestly such a good conversation. We started with Clay and outbound, and somehow ended up talking about Claude Code, databases, APIs, AI agents, workflows, CRM and how all of this fits together. One thing I really liked was the idea of actually understanding the systems you’re building. If you’re using Claude Code, you don’t necessarily need to know how to code everything yourself. But you should be able to look at what it built and understand what’s happening. We also got into the whole AI agents conversation. What actually needs to be an agent? What can just be a workflow? That part was really interesting. The coding agent discussion was probably my favourite. Instead of throwing a huge CSV at Claude Code and saying “research all of this”, you can structure the work using scripts, APIs and different models for different parts of the process. You basically give the agent a job to orchestrate rather than asking it to do everything. That was a really useful way for me to think about it. Really enjoyed this one. 😊 Thanks Kushagra T. for taking the time to do this with us, and Yogesh Jaiswal for putting the whole session together. 🙌

Claudeforce Lands: What Salesforce’s Anthropic Deal Means for Sellers Salesforce and Anthropic's new Claudeforce plugin lets reps work live CRM data inside Claude. Here's what the Q2 earnings surge tells us about where AI monetisation is heading. Read the full article → https://lnkd.in/g_m4H45y #CRMDaily #CRMNews #GTM #RevOps #Salesforce #HubSpot #SalesTech

How are people using AI in day-to-day work — and what are they building with it? 🇨🇦 Join Canadian AI Literacy and Innovation Foundation September AI Meetup for a practical Claude workshop about agentic workflow, project management and GTM. There will be CAILIF member demos, free pizza, and a networking session. Wednesday, September 2, 6:30–8:30 PM 4185-C Still Creek Dr, Burnaby Open to everyone! If you are on the waitlist, ping me to get you in. RSVP: https://luma.com/neigsgpw #AI #ClaudeAI #VancouverTech #CAILIF #AIMeetup #CanadaAI #ALLINAI

If I were building a GTM engineering system in 2026, these are the tools I'd connect: - HubSpot / Salesforce → closed-won + closed-lost data to understand what actually converts - Clay + ChatGPT + Claude → analyse customer data and build a more accurate ICP - Clay + LinkedIn Sales Navigator + Ocean.io → map the TAM instead of relying on one database - Perplexity + LinkedIn + company websites → research accounts and add useful context - Clay + LinkedIn Sales Navigator → enrich accounts and identify the right people - HubSpot / Salesforce / Attio → keep the CRM as the system of record - Clay + AI → qualify, score and route accounts based on ICP + situation + signals - Awareness scoring → understand whether an account is unaware, aware, researching, considering or buying - Outbound + inbound + ABM → activate the right GTM play for the account and situation - HubSpot / Salesforce → connect meetings and sales activity back to closed-won revenue Every tool has a specific role. And every layer should: - Improve the data - Improve the targeting - Improve the context - Improve the next decision (and no, adding 15 more tools doesn't magically create a GTM system 😂) The goal isn't to have the biggest stack. It's to make the stack work together: customer data → ICP → TAM → research → qualification → signals → GTM play → revenue What would you add to this system?

Chaque matin, Claude me sort 50 prospects ultra-chauds. Voici les signaux que je lui donne. Je n'achète plus de listes. Je ne scrolle plus LinkedIn au hasard. Je ne pars plus d'un fichier froid de 2000 lignes dont 1980 ne bougeront jamais. Chaque matin, j'ouvre, et j'ai 50 boîtes qui viennent de faire quelque chose qui dit « on a un besoin, maintenant ». Les signaux que je surveille : → Elle recrute un commercial. Elle veut vendre plus, mais elle n'a encore aucun système pour ça. → Un directeur commercial vient d'arriver. Il doit faire ses preuves vite, il cherche des leviers. → Elle change de CRM. Elle repense toute son organisation commerciale. Le moment parfait. → Un dirigeant qui s'intéresse au GTM sur LinkedIn. Il cherche déjà. Il est en pleine réflexion. → Une phase de croissance. Elle grossit, sa prospection artisanale ne suit plus. → Du turn-over dans les équipes sales. À chaque départ, elle perd son savoir. Elle a besoin d'un système qui ne parte pas avec les gens. Un signal seul, c'est une piste. Deux signaux croisés sur la même boîte, c'est un prospect chaud. Pendant que les autres appellent 200 personnes qui n'ont rien demandé, moi j'appelle 50 boîtes qui viennent littéralement de lever la main. Le volume n'a jamais été le problème. Le timing, si. Tes signaux à toi seront différents des miens. J'ai mis au clair la méthode pour les définir, les repérer et les croiser. Tu veux la méthode ? Commente « SIGNAUX » et je te l'envoie.

This used to be two days of work for a GTM engineer: "Find me 2000 US SaaS companies that raised funds, get their CEO, Head of Sales and CRO, find their emails and phone numbers, and push them into my Instantly campaign." Now it's one sentence typed into Claude Code via the ColdIQ MCP. A few minutes later, the campaign is loaded. Here's what's underneath it. 40+ data providers behind one API key. You describe what you want. It picks based on fit, cost and accuracy, then waterfalls to the next one when the first comes back empty. No dashboards. No 40+ logins. No cleaning 40+ response formats. And that's 1 use case out of a million.

Mitä tapahtuu kun testaa Claude Codella optimoida vanhaa WP-sivustoa? Tulos 100% tehokkuudesta, esteettömyydestä, parhaista käytännöistä ja hakukoneoptimoinista + 2/2 agenttiselauksesta 🫣 Näitten kanssa tuli aiemmin painuttua tunteja, nyt tämä hoitui noin kolmella testikierroksella. Melko huikeeta, en tiedä kyllä kuinka paljon tästä nykypäivänä on apua, mutta eipä tästä nyt ainakaan haittaa ole. Perinteisesti kuvat olivat sivun raskain osa, joten näihin tehtiin WebP-muunnos. Tämän jälkeen kaivoimme syyt siihen, miksi mobiilipisteet jäivät vajaaksi. Tähän löytyi kaksi syytä: ➡️ WordPress lisää lazyload-kuviin sizes="auto". Se sai selaimen valitsemaan ensin liian ison variantin ja vaihtamaan oikeaan vasta asettelun jälkeen, sama kuva latautui kahdesti. ➡️ Hero-lohkon geometria mitattuna: puhelimessa tekstipaneeli peittää kuvan koko pinta-alalta ja on 96-prosenttisen läpinäkymätön. Valokuvasta jää läpi vain hienovarainen sävy. Tässä voitiin käyttää sitten pienempää kuvaa mobiilissa. Lopputulos: etusivun kuvakuorma mobiilissa keveni 81 % ja LCP-elementin tiedosto 79 %. Ulkoasu ei muuttunut pikseliäkään. Deskari säilyi täydellä tarkkuudella, koska siellä kuva oikeasti näkyy. Kiinnostavinta oli työtapa. Claude Code ei arvannut vaan mittasi ja korjasi: avasi sivun selaimessa eri näyttöleveyksillä, luki verkkopyynnöt ja laski, paljonko alue oikeasti peittää. Näähän on ihan tiedossa olevia juttuja mutta kiva oli ettei itte tarttenut koskea koodiin antoi CC:n tehdä vaan hommat. Yksi huomio, sivustolla ei ole GTM:ää käytössä vielä, joten sen kanssa ei tarvinnut tapella, mutta tästä on hyvä jatkaa. Suosittelen testaamaan 😉

What if your brand is no longer competing for a Google ranking — but for a place in the answer generated by AI? I recently came across a fascinating episode of The Core Report Weekend Edition, where Govindraj Ethiraj speaks with Kirthiga Reddy, Co-founder & CEO of OptimizeGEO.ai, about how AI is fundamentally changing search, marketing and brand discovery. One idea particularly stood out to me: In the age of AI, being visible is no longer enough. You need to be recommended. As consumers increasingly move from traditional Google searches to asking ChatGPT, Gemini, Claude and Perplexity what to buy, which brands to trust and what best fits their needs, the rules of brand discovery are changing. This creates a new discipline: Generative Engine Optimization (GEO). Unlike traditional SEO, GEO isn't simply about ranking on a search results page. It's about ensuring that AI systems: • Know your brand • Understand what you offer • Have consistent and accurate information about you • Consider you relevant to a particular need • Associate your brand with trust and positive sentiment • Ultimately, include you in the answer And there is a bigger implication for marketers. AI agents may increasingly become the new gatekeepers between consumers and brands. If an AI agent researches products, compares alternatives and eventually makes a purchasing recommendation, companies will need to learn how to market not only to people, but also to the AI systems influencing those people's decisions. That could fundamentally reshape customer acquisition. The shift from SEO → GEO → Agentic Commerce may be one of the most important changes happening in digital marketing right now. Worth listening to for anyone involved in marketing, sales, digital strategy, brand building or AI-led GTM. 🎙️ Credit: The Core Report Weekend Edition by Govindraj Ethiraj, featuring Kirthiga Reddy, Co-founder & CEO, OptimizeGEO.ai. #AI #GenerativeAI #GEO #GenerativeEngineOptimization #SEO #AISearch #Marketing #DigitalMarketing #BrandStrategy #CustomerAcquisition #AgenticCommerce #AIGTM #FutureOfMarketing #ChatGPT #ArtificialIntelligence

Criadores não precisam apenas de comandos para escrever. Precisam de um sistema que transforme conhecimento em conteúdo, distribuição, conversão e aprendizado. Use os 18 comandos por objetivo: CRIAÇÃO /GANCHO: aberturas que prendem atenção. /CARROSSEL: transforma uma ideia em sequência. /THREAD: cria narrativa em publicações conectadas. /CAPTION: melhora legendas. /REWRITE: reescreve conteúdo fraco. APRENDIZADO /ELI10: explica como para uma criança de 10 anos. /PRIMER: cria introdução estruturada. /QUIZ: testa compreensão. /ANALOGIA: simplifica conceitos. /FIRSTPRINCIPLES: decompõe até os fundamentos. ESTRATÉGIA /COMPARE: compara opções com critérios. /AUDIT: encontra falhas e oportunidades. /SCALE: identifica alavancas de receita. /SYSTEM: transforma tarefa em sistema. /AUTOMATION: desenha fluxo automatizado. /DECIDE: estrutura uma decisão. /GTM: cria plano de entrada no mercado. /CTA: melhora a próxima ação. COMO USAR: 1. Escolha o resultado desejado. 2. Cole dados, público, canal e objetivo. 3. Defina tom, tamanho e restrições. 4. Execute um comando por etapa. 5. Valide precisão, utilidade e aderência à sua voz. 6. Salve a sequência aprovada em uma Skill. PROMPT PARA CRIAR O SISTEMA: “Transforme os 18 comandos em fluxos reutilizáveis para meu negócio. Para cada comando, defina entradas, perguntas obrigatórias, execução, saída, checklist e conexão com o próximo comando. Contexto: [NEGÓCIO]. Público: [PÚBLICO].” O comando não substitui estratégia. Ele transforma uma estratégia já definida em operação repetível. Siga Lucíola Coelho, o perfil LinkedIn #1 no Mundo 2025 e 2026 + IA. #luciolacoelho #aiia #claude #criadores #conteúdo #skills #engenhariadeprompt #genai

Marketing in 2026: The System Shift Your GTM stack probably isn't missing another tool. It's missing a SYSTEM. I researched Reddit, Product Hunt & top AI reports — yeh 4 shifts har marketer ko samajhna zaroori hai: 1. AI Stack: ChatGPT, Claude, Gemini, Perplexity + Autonomous Agents. Tools replace tasks, and systems replace teams. 2. SEO → GEO/AEO: Goal is no longer Rank #1. Goal is to be cited in AI answers. Entity & Trust is the new backlink. 3. Performance: First-party data + predictive + privacy-first + omnichannel. No more manual bidding. 4. UX/UI: AI-First, Bento Grid, Zero UI. Interface khud soch ke action le raha hai. Old playbook dead hai. 2026 mein jeetne wale woh hain jo tools nahi, systems banate hain. Aap kis shift pe kaam kar rahe ho? Comment karo. #AIMarketing #GEO #AEO #PerformanceMarketing #UXUI #DigitalMarketing2026 #RizwanSaeed

Marketing in 2026: The System Shift Your GTM stack probably isn't missing another tool. It's missing a SYSTEM. I researched Reddit, Product Hunt & top AI reports — yeh 4 shifts har marketer ko samajhna zaroori hai: 1. AI Stack: ChatGPT, Claude, Gemini, Perplexity + Autonomous Agents. Tools replace tasks, and systems replace teams. 2. SEO → GEO/AEO: Goal is no longer Rank #1. Goal is to be cited in AI answers. Entity & Trust is the new backlink. 3. Performance: First-party data + predictive + privacy-first + omnichannel. No more manual bidding. 4. UX/UI: AI-First, Bento Grid, Zero UI. Interface khud soch ke action le raha hai. Old playbook dead hai. 2026 mein jeetne wale woh hain jo tools nahi, systems banate hain. Aap kis shift pe kaam kar rahe ho? Comment karo. #AIMarketing #GEO #AEO #PerformanceMarketing #UXUI #DigitalMarketing2026 #RizwanSaeed

🌎 Vagas 100% remoto 💰 Pagamento em dólar 📩 Candidatura únicamente pelos links ✅️ AI Engineer - Python, LLM, AI Agents 2.1k-3.5k USD/month | 3+ years of exp. | B2 English https://lnkd.in/dZVhYWEq ✅️ Senior GTM Engineer - GTM processes, Salesforce, HubSpot 3.5k-5.5k USD/month | 5+ years of exp. | B2 English https://lnkd.in/dUJRiemQ ✅️ Product Engineer - Python, React.js, Node.js 4.9k-5.6k USD/month | 5 to 8 years of exp. | C1 English https://lnkd.in/dXmXsjVd ✅️ LiDAR Specialist - LiDAR, GIS, ArcGis 1.4k-3.5k USD/month | 5+ years of exp. | B2 English https://lnkd.in/dECcS2M7 ✅️ Senior Database Administrator and Developer - Snowflake, SQL Server 3.5k-5k USD/month | 7+ years of exp. | C1 English https://lnkd.in/dGW-tBZA ✅️ Full-stack Engineer - Terraform, PostgreSQL, React.js 4.55k-5.6k USD/month | 5+ years of exp. | C1 English https://lnkd.in/dyDQVA8x ✅️ Full-stack Engineer - UX Design, React.js, Node.js 4.3k-6k USD/month | 4+ years of exp. | C1 English https://lnkd.in/dwE9ngFE ✅️ Growth Marketer - AI Technologies, Claude 2.1k-3.5k USD/month | 4 to 6 years of exp. | C1 English https://lnkd.in/dK_pSmNG ✅️ Platform Engineer - Kubernetes, Terraform, CI/CD 3.5k-5k USD/month | 5+ years of exp. | C1 English https://lnkd.in/d8p8J76F ✅️ Product Engineer - Java, Kotlin, Golang 3.5k-5k USD/month | 5+ years of exp. | C1 English https://lnkd.in/dSrjiHRm

I’m evaluating Perplexity Computer through a founder lens. What I need is not another AI tool for writing drafts. I need leverage: help turning messy questions into decision-ready work without personally doing every hour of research, synthesis, and coordination. I’m hoping it can help with: Competitive and market research - Synthesizing customer and sales feedback - GTM and product-planning first drafts - Decision memos, operating updates, and next-step recommendations. My expectation is not that it replaces my judgment. It is that it compresses: Question → research → insight → decision → action (most important) But I’m also looking for the gotchas, tips and tricks, where it works, where it doesn't etc. Where does Perplexity Computer genuinely create leverage versus just produce polished-looking output? What workflows have other founders successfully delegated? Where do accuracy, privacy, permissions, or quality-control issues show up? And what alternatives should I evaluate alongside it ? #Claude, #ChatGPT, #Gemini, custom agent workflows, or vertical AI tools? I’d love to hear what is actually working for other founders. #PerplexityComputer #AIAgents #FounderLeverage #AI #Startups @ADHD the link Unlockt

〈a16z 深度對談：Cursor——讓 AI 寫代碼成為必然的公司〉 [本文由 AI 依據 YouTube 影片生成] 頻道：a16z ｜ 日期：2026-08-27 ｜ 觀看：1582 a16z 合夥人 Martine、Matt 與成長基金合夥人 Sarah 在這次對談中，完整複盤了他們從 2024 年初開始、橫跨 Series A 到後續多輪投資的 Cursor（Anysphere）投資全紀錄。三位投資人不僅回顧了當時 AI 程式碼生成領域的競爭版圖，更深入剖析了 Cursor 團隊如何在微軟 Copilot、Anthropic Claude Code 等巨頭圍堵下，憑藉極致的產品專注、反直覺的策略抉擇、以及近乎偏執的文化與招聘標準，實現了「垂直起飛」般的成長。 回溯到 2023 年底、2024 年初，主流模型為 GPT-4、Claude 3 與 Llama 3，程式碼生成領域的共識是：微軟擁有 VS Code、GitHub Copilot、OpenAI 權重與企業銷售網絡，幾乎已經鎖定勝局。當時創業團隊大致分為兩派：一派認為必須訓練專屬程式碼基座模型，另一派則專注於 Agent 架構；Cursor 創辦人 Michael Truell、Aman Sanger、Sualeh Asif 選擇了一條中間道路——他們深受「苦澀教訓」影響，堅信當下不需與 Anthropic、OpenAI 爭奪模型訓練，核心在於「人與模型之間的介面」。Michael 早期反覆強調：「未來的寫代碼會像偽代碼」，程式設計師只需給出最小規約的自然語言意圖，模型即可實作。這一洞見貫穿了 Cursor 從 IDE 插件、Agent 平台，再到自研模型平台的演進。 a16z 早期之所以下注，關鍵在於團隊的「異常專注」與「頂尖工程師的腳踏實地」。Martine 回憶，OpenAI、Midjourney、Replicate 等一線 AI 公司的工程師已大量使用並訂閱 Cursor；而在向 GP 條陳時，Michael 將 90% 的時間用於對各種 VC 想像出的副業說「不」——不做企業版、不做插件、不分心。Matt 指出，這種清晰源於他們對「產品型公司」的自我定義：若真信仰產品，就不會做插件（變成別人產品的一部分），也不會過早做企業銷售（產品本身才是最強的去市場動作）。這種決策一致性，在當時充滿「拼湊式策略」的競爭者中極為罕見。 成長基金介入時面臨極大認知衝突：Cursor 從 400 萬美元 ARR 衝到 5000 萬美元僅用幾個月，卻完全違背教科書——自助式增長沒有放緩、不需要在 2500 萬 ARR 時引入銷售主管、毛利率與商業模式在 X 平台上被投資人批評。Sarah 與 David George 最終決定「拋棄傳統成長模型的假設」，押注於團隊對市場動態的即時反應能力。競爭面上，Copilot 是首個巨頭威脅，隨後 Windsurf 在 YC 圈層崛起、Cognition 推出 Devin Agent、Anthropic 於 2025 年 5 月發布 Claude Code。面對這輪輪「輪換陣容」的強敵，Michael 的回應極具啟發性：「我們鎖定的是全球最大市場， formidable competitors 永遠存在，這不會嚇倒我們。」這種「謙遜中帶著膽識」的態度，讓 Sarah 聯想到貝佐斯與伊隆·馬斯克面對質疑時的堅韌。 確實有過幾次「Oh shit」時刻——Claude Opus 4.5 與 Mythos 預發布版本釋出時，模型能力躍升速度遠超預期。但 Cursor 團隊展現了類似奈飛瑞德·哈斯汀斯式的自我吞噬勇氣：兩年內從 IDE、到 Agent 平台、再到模型平台，主動淘汰自身舊有優勢（如 Karpathy 曾推崇的 Tab 鍵補全已非核心）。對於外界對毛利率與商業模式的質疑，Martine 引用互聯網早期歷史反駁：「技術變革永遠先於商業模式；當你擁有無限需求與增長時，利潤只是時間問題。」Cursor 以「先拿用戶、再拿數據與知識、最後反過來練模型」的後門策略，雖曾挨罵，如今顯然正在兌現。 企業級銷售的轉型同樣打破教科書。早期 Michael 甚至拒絕建立銷售團隊，僅設立 enterprise@cursor.com 郵箱；但當意識到大模型廠商將補貼個人版、利潤集中在企業端時，他們以創業初期 40% 時間用於招聘的強度，打造了「歷史上最快組建的銷售團隊」。Jordan、Roman 主導的 GTM 招聘沿用工程端的「Epoch 分析法」：鎖定過去特定階段表現最強的前 10 家公司、前 10 個團隊、第 1、2 號王牌 AE，層層背調至老闆的老闆的老闆。創始 AE 需具備在不確定性中作戰的能力，而非單純跑流程。結果：Cursor 在極短時間內滲透超過 50% 的財富 500 強企業。 文化層面，三位投資人一致觀察到：Cursor 團隊不是「迷戀工作」，而是「真心熱愛寫代碼與系統」。這份純粹延伸至辦公空間設計（北歐風、脫鞋換拖鞋、二手家具營造居家感）、行銷聲音（拒絕噴發式行銷、堅持品牌調性）、人員決策（不合即快速切割）、以及 M&A 整合。早期以人才併購為主（如 Koala 創辦人 Tito、Adam Ward），後期 Graphite 併購更具戰略意義；儘管操作複雜度極高，Cursor 卻因文化吸納力極強，將前創辦人無縫融入並釋放杠桿。Martine 評價這或許是「歷史上最專注、卻又完成最多 M&A 的公司」。 對談尾聲，Matt 與 Sarah 將 Cursor 與 SpaceX 並列：同樣面對共識性的「死亡預言」、同樣在看似不可能的競爭中贏下巨大市場、同樣擁有極速迭代、硬核決策與工程文化。若未來併購成真，這將是技術、願景與文化三維度極罕見的完美契合。 --- 關鍵概念 Cursor, Anysphere, a16z, Andreessen Horowitz, Michael Truell, Aman Sanger, Sualeh Asif, Andrej Karpathy, Lex Fridman, Microsoft, GitHub Copilot, VS Code, Anthropic, Claude, Claude Code, OpenAI, Midjourney, Replicate, Windsurf, Codeium, Cognition, Devin, Elon Musk, SpaceX, Starlink, Reed Hastings, Netflix, David George, John Schulman, Jordan, Roman, Bri, Tito, ……（未完） 標題：How Cursor Built One of AI’s Fastest-Growing Companies 摘要：https://lnkd.in/ehzuVtE4 來源：https://lnkd.in/e4d8ERU3

I mapped out all 50 skills inside a Claude system built specifically for startups and scale-ups. Here's the full breakdown, category by category, because "50 AI skills" means nothing without seeing what's actually in it. Market Intelligence Market Opportunity Mapper, Competitor Intelligence Analyst, Category Researcher, Market Trend Analyzer, Growth Opportunity Prioritizer. ICP & Customer Research ICP Builder, Buyer Pain-Point Miner, Buying Trigger Mapper, Buying Committee Mapper, Customer Interview Synthesizer. Positioning & Offers Positioning Architect, Value Proposition Builder, Offer Designer, Messaging Hierarchy Builder, Objection Intelligence Analyst. Go-to-Market Strategy GTM Architect, Channel Selection Engine, GTM Experiment Designer, Launch Strategist, GTM Bottleneck Diagnostician. Demand Generation Demand Generation Planner, Content-to-Pipeline Strategist, Lead Magnet Creator, Campaign Architect, Distribution Strategist. Prospecting & Buying Intent High-Intent Prospect Prioritizer, Buying Signal Interpreter, Account Intelligence Researcher, Prospect Researcher, Prospect Prioritization Engine. Outbound & Sales Conversations Outreach Angle Selector, Signal-Based Outreach Writer, Discovery Call Prep, Discovery Call Analyzer, Sales Follow-Up Builder. Pipeline & Conversion Lead Qualification Engine, Pipeline Diagnostician, Deal Risk Analyzer, Proposal Strategist, Conversion Friction Analyzer. Customer Growth Customer Onboarding Architect, Customer Health Analyzer, Churn Diagnostician, Expansion Opportunity Finder, Voice-of-Customer Miner. Growth Strategy & Optimization Growth Experiment Designer, Funnel Diagnostician, Growth Metrics Architect, Growth Review Analyst, Growth Priority Engine. That's the entire system, start to finish: understanding the market, finding the right customers, positioning the offer, building the GTM plan, generating demand, prioritizing the right prospects, running better sales conversations, tightening the pipeline, growing existing customers, and continuously optimizing growth. Nothing in that list replaces a team. It replaces the hours a team spends starting each of these from a blank page. Which category above is the weakest link in your current process? 👇

Team-led content is cheat code for B2B growth in 2026. Companies are generating millions with exactly this strategy ↓ (Save it before it gets lost in the feed.) Most companies still create content like this: Open a blank AI chat. Write a prompt. Fix the generic draft. Repeat everything tomorrow. Here is how best teams operate. They create a folder which claude or codex can access with: 𝟭. 𝗕𝘂𝘀𝗶𝗻𝗲𝘀𝘀 𝗯𝗿𝗮𝗶𝗻 ICP, positioning, messaging, offers, proof, objections, customer language, and the founder's point of view. 𝟮. 𝗥𝗲𝗳𝗲𝗿𝗲𝗻𝗰𝗲 𝗹𝗶𝗯𝗿𝗮𝗿𝘆 The best posts, hooks, infographics, carousels, and images are saved as quality benchmarks. The system studies the patterns without copying the work. 𝟯. 𝗣𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻 𝗽𝗹𝗮𝘆𝗯𝗼𝗼𝗸𝘀 Step-by-step instructions for posts, carousels, infographics, reviews, and repurposing. HTML templates keep every visual consistent and editable. 𝟰. 𝗜𝗱𝗲𝗮 𝗲𝗻𝗴𝗶𝗻𝗲 Reddit threads reveal pain. X shows emerging conversations. Competitors reveal crowded angles. Sales calls provide buyer language, objections, and proof. 𝟱. 𝗥𝗲𝘃𝗶𝗲𝘄 𝗹𝗼𝗼𝗽 Every draft is checked against the ICP, positioning, voice, proof, hook quality, and pipeline outcome before it ships. Now a team member does not need to guess what the founder means. They open the workspace, follow the playbook, and improve the shared system with every post. AI handles execution. The team's accumulated context protects the quality. What is your content creation workflow with AI? ♻️ Repost to help your founder friend nail GTM ➕ Follow Kanchan Bhatta for GTM systems My Tool Stack: Claude Code Cursor Google Workspace Reddit, Inc. X Higgsfield AI #GTM #B2BSaaS #FounderLedGrowth

I built an entire GTM department inside Claude. and I’m giving it away for free ↓ it's a complete GTM system where every function works together. I split it into 6 departments: 1/ GTM Intelligence define your ICP, map buyers, research competitors, and analyze the voice of your customers 2/ Positioning & Offer build your positioning, sharpen your offer, create messaging, and turn proof into something that sells 3/ Demand & Content create your content strategy, write for every channel, build landing pages, and generate lead magnets 4/ Outbound find accounts, detect buying signals, score prospects, research leads, write sequences, and handle replies. 5/ Sales prepare discovery calls, handle objections, build proposals, and inspect deals before they stall. 6/ Revenue & Retention improve onboarding, identify expansion opportunities, and analyze what’s actually driving revenue. the best part is very department works from the same shared GTM brief so claude already knows your product, ICP, positioning, competitors, and customers you don't have to: - start from scratch - re-explain your business - try disconnected prompts just one GTM department working from the same context want the full Claude GTM Department? giving it away for free below 👇

🚀 Hiring | Sr. Paid Marketing Analyst | Google Ads | Meta | LinkedIn | Pune (On-site) GO MO Group is Hiring!!! If you have strong hands-on experience in Google Ads, Meta Ads, LinkedIn Ads, and performance marketing, we'd love to hear from you! 📩 Send your resume to: pratiksha.j@gomogroup.com Or simply drop me a DM. 💻 Role: Sr. Paid Marketing Analyst 🧑‍💻 Experience: 2–3+ Years 📍 Location: Kalyani Nagar, Pune 🏢 Type: On-site | Permanent We're Looking For i) 2–3+ years of experience in paid media/performance marketing. ii) Strong hands-on experience with Google Ads (mandatory), Meta Ads & LinkedIn Ads. iii) Experience with GA4, GTM, Looker Studio, and performance analytics. iv) Strong understanding of bidding strategies, attribution, CPC, CTR, CPA & ROAS. v) Data-driven approach with experience managing medium to large budgets. vi) Experience using AI tools such as ChatGPT/Claude for campaign optimization. vii) B2B client-facing experience is a plus. Know someone who'd be a great fit? Please like, comment, or share this post to help us reach the right talent! #Hiring #NowHiring #PaidMarketing #PerformanceMarketing #GoogleAds #MetaAds #LinkedInAds #DigitalMarketing #MarketingJobs #PuneJobs #PuneHiring #HiringInPune #PaidMedia

GojiberryAI (YC) is a $4M ARR company with just 8 people and half of their pipeline comes from the 3 founders posts. Roman : Founder of GojiberryAI. He has 19K followers on X. Consistently shares playbooks, experiments, and real numbers. His content mix: → Educational (Teaches people what’s actually worked, not generic info) → Product experiments → Lead magnets → Personal posts → Behind the scenes (ad spend, demos/week, team LinkedIns running campaigns) How he converts: teach the whole playbook → send the resource → drop the 7-day trial. Pierre-Eliott: CEO of GojiberryAI, 9.2K followers on X. Mostly shares behind the scenes content, hiring posts. His content mix: → Long tutorials (Claude + high-intent leads, deliverability, full outbound systems) → Company math ($4M ARR, team of 8, what they’re rebuilding) → SF / YC / “we just hired” proof → Actual job posts How he converts: authority first. Video + playbook + “try it free.” Dylan: He is the co-founder. 1.2K followers on X. His content mix: → Stage-by-stage ARR posts ($0 → $6k → $25k → $75k → $150k+) → Intent-based outbound frameworks → Experiment logs (10 ICPs × 136 sequences) → Meme posts sometimes → Customer stories How he converts: he makes the strategy feel obvious, then positions Gojiberry as the system that runs it. They also run other channels like B2B influencer marketing on LinkedIn, ads, cold emails. But all others work 2-3x better because of the founder-led content first. P.S. I predict the founder will become the entire GTM of a company within a few years. Drop your thoughts below.

Earlier this year I said out loud that my 2026 manifestation was to build a one-of-a-kind recruitment founders' retreat. ⠀ Last week, we made it happen. ⠀ Three days, ten agency founders, one magnificent villa by the ocean. ⠀ You decide what you want to achieve, you say it publicly before you have any idea how you will pull it off, and the decision starts pulling the pieces toward you. ⠀ We've built a close-knit community at RecruiterGTM where I can take my ideas to validate them. I ran a poll in April about organizing this event and had enough interest to pursue the idea and now we're definitely turning this into an annual thing. ⠀ Over three days I personally set up a Claude Ops Manager in each of their environments and deployed their ContentOS engine with them. ⠀ The systems were only half of it. ⠀ Most of these founders had only met each other on video calls and were equally excited to spend some quality time together and it was a great opportunity for me to connect with them more closely. ⠀ We´re already planning the next one for early summer in 2027, and deciding between Portugal, Italy or Turkiye. ⠀ If you run a recruitment agency and want first shout on a seat, send me a DM. ⠀ —— ⠀ I'm Reyhan Khan, I post about how we help recruitment businesses install GTM Systems & Claude Code allowing their offers to stand out. If you want to follow the journey more closely, the newsletter link is in the first comment.

GTM Engineers, you have to understand how everything works, its your job! By now, we all heard of ChatGPT, Gemini, Claude. But have you used Codex, Claude Code, or Cursor. Did you know Cursor is owned by SpaceX and because of that OpenAI is leaving in November. Do you understand Open Source, have you heard of GLM, are you familiar with open weights. Can you install a CLI or MCP? Its a lot, I know. But as a GTM, you are expected now to be just as much an AI Engineer at you are a GTM Engineer. Are you testing Grok and Devin. Did you know Grok has its own computer, yet most work you do in Codex is on your local machine. Your boss will expect you to know how data will be protected, what is being shared and most importantly, what does everything cost. You need to be able to explain why you need what you need to do your job. Your business has a budget, and unless you work for a unicorn, its not unlimited!

🏢 ¿Y si en lugar de un único asistente empresarial tuviéramos un equipo ejecutivo completo de agentes especializados? Estuve analizando OpenExecutive, un proyecto open source de Sente Labs que implementa un virtual executive team sobre Claude. 🔗 GitHub: https://lnkd.in/dqA5-e2T La arquitectura es: Usuario ↓ Executive Orchestrator ↓ Especialistas ↓ RAG + memoria empresarial ↓ Síntesis ↓ Una única respuesta ejecutiva El sistema trabaja con 8 roles: 🎯 Estrategia 💰 Finanzas 👥 People / HR ⚖️ Legal ⚙️ Operaciones 📣 Marketing 📦 Producto 📊 Board Communications Pero el usuario no conversa con ocho bots diferentes. Los especialistas trabajan por detrás y el sistema consolida sus análisis en una sola respuesta. Multi-agent por dentro. Una única interfaz por fuera. Cada agente puede recuperar contexto desde: 📚 conocimiento empresarial incorporado 📄 documentos propios de la compañía El RAG utiliza ChromaDB y permite incorporar pitch decks, modelos financieros, estrategia y documentación interna. Además tiene una capa distinta de: 🧠 Episodic Memory Después de las conversaciones puede extraer: • decisiones • iniciativas • recomendaciones y almacenarlas en SQLite para recuperarlas en sesiones futuras. Eso cambia el modelo: Chat → respuesta → olvidar por: Conversación → decisión → memoria → seguimiento También incorpora comportamiento proactivo. Un scheduler puede detectar: ⏰ acciones pendientes 🔔 compromisos 📌 iniciativas abiertas y generar recordatorios sin esperar una nueva consulta. El repositorio incluye workflows para escenarios como: • Annual / Quarterly Planning • Board Prep • Fundraising • M&A • Pricing • GTM • Product Strategy • Risk Management • Investor Updates • Morning Brief Además puede operar mediante: 🌐 Web 💬 Slack 📧 Email ✈️ Telegram 🎮 Discord ⌨️ CLI El stack incluye: 🧠 Claude 🐍 FastAPI ⚛️ Next.js 15 🗄️ ChromaDB 💾 SQLite 🐳 Docker Desde el punto de vista de arquitectura, lo interesante para mí es esto: Un sistema empresarial de IA quizá no debería ser: un chatbot que sabe de todo. Podría ser: Orquestador + especialistas + memoria + conocimiento corporativo + workflows + automatización Eso empieza a parecerse más a una capa operativa de inteligencia alrededor de una organización. ⚠️ Naturalmente, agentes de finanzas, legal, RRHH o estrategia deben apoyar decisiones, no sustituir la responsabilidad y gobernanza humana. El proyecto está todavía en una etapa temprana: v0.1.0. Open source bajo Apache License 2.0. 🛠️ #ArtificialIntelligence #AIAgents #AgenticAI #AIEngineering #EnterpriseAI #MultiAgentSystems

The future of AI at work is not better chat. It is moving from a request to a result. This does not mean replacing human judgment. It means redesigning repetitive workflows so people can focus on strategy, creativity, decisions and execution. One recent example is our Friday Fintech Intelligence workflow. For anyone familiar with Claude Cowork, ChatGPT Work is the closest OpenAI analogue: a workspace where an agent can complete substantial tasks using context, files and connected business tools. The business need was straightforward: Every Friday, identify the most important developments across African fintech, consumer credit, GTM, lifecycle growth, AI and competitive threats. But the objective was not to produce another research summary. The workflow needed to: - Conduct and validate the research - Prioritize what matters - Translate signals into strategic implications - Recommend real business actions - Use the company's presentation template - Produce mobile and desktop reports - Draft the executive email - Prepare distribution across management teams and companies - Track what happens to the recommended actions The prompt was developed iteratively. - First, I defined the research scope. - Then I added the strategic decision lens. - Next came the company template, management-ready formats, source validation, design quality checks and email preparation. - Finally, I added the operational layer: actions, owners, status and follow-through. The result is a simple agentic operating loop: Research -> intelligence -> management communication -> action -> tracking -> next review. - A scheduled task brings the agent back every Friday with the existing context. - Deep Research finds and validates the evidence. - Skills preserve the working method. - Plugins and connectors provide access to the necessary tools. ChatGPT Work coordinates the process. The human remains in control of important decisions, approvals and external communication. This is what working with agents looks like in real life. Not one impressive answer. A repeatable system connecting research, company knowledge, management communication and execution. This is only one relatively simple automation within a much broader transformation. As AI adoption across our marketing operations moves beyond 70%, the next challenge is not finding more tools. It is connecting the tools into accountable workflows that produce measurable business outcomes. #ChatGPTWork #AgenticAI #MarketingOperations #FutureOfWork #Fintech

Was excited and grateful to join the RIB team Akshaya (AK) Dash and Nishanth Patil — for a working session with Anthropic on how teams are using AI across software development. Chase Romano gave us a hands on master class! Thank you! I took a bunch of notes during the conversation and organized them into two buckets: 𝟭. 𝗧𝗼𝗼𝗹𝗶𝗻𝗴 — 𝗛𝗼𝘄 𝗱𝗼 𝘄𝗲 𝘂𝘀𝗲 𝗔𝗜 𝗲𝗳𝗳𝗲𝗰𝘁𝗶𝘃𝗲𝗹𝘆? In this instance, the tooling conversation was specifically around Claude. 𝟮. 𝗣𝗿𝗼𝗰𝗲𝘀𝘀 — 𝗛𝗼𝘄 𝗱𝗼𝗲𝘀 𝗔𝗜 𝗰𝗵𝗮𝗻𝗴𝗲 𝘁𝗵𝗲 𝘄𝗮𝘆 𝘄𝗲 𝗯𝘂𝗶𝗹𝗱 𝗮𝗻𝗱 𝗱𝗲𝗹𝗶𝘃𝗲𝗿 𝘀𝗼𝗳𝘁𝘄𝗮𝗿𝗲? This is the broader AI-DLC conversation and is largely tool-agnostic. Here are my 𝗕𝘂𝗶𝗹𝗱𝗲𝗿 𝗡𝗼𝘁𝗲𝘀 from the conversation, many of us maybe doing some of these already ! 𝗖𝗹𝗮𝘂𝗱𝗲 𝗮𝘀 𝗮 𝗧𝗼𝗼𝗹 - 𝗛𝗮𝗿𝗻𝗲𝘀𝘀 𝗮𝗻𝗱 𝗠𝗼𝗱𝗲𝗹 • Keep 𝙲𝙻𝙰𝚄𝙳𝙴.𝚖𝚍 short and clear — use it to provide persistent repo-level context without turning it into another documentation dump. • Use 𝘀𝘂𝗯-𝗮𝗴𝗲𝗻𝘁𝘀 𝗳𝗼𝗿 𝗳𝗼𝗰𝘂𝘀𝗲𝗱 𝘄𝗼𝗿𝗸 to preserve the main context, • Route work deliberately across models — frontier models for planning and complex reasoning (Fable, Opus) ; smaller models for execution and triage (Sonnet, Haiku) • 𝗖𝗼𝗻𝗻𝗲𝗰𝘁 Claude to 𝘄𝗵𝗲𝗿𝗲 𝘄𝗼𝗿𝗸 𝗮𝗹𝗿𝗲𝗮𝗱𝘆 𝗵𝗮𝗽𝗽𝗲𝗻𝘀 — repositories, documentation, issue tracking, and collaboration tools (via MCP) • Plug Claude into CICD process - through GitHub Actions for instance for automatic Security Checks as an example 𝗔𝗴𝗲𝗻𝘁𝗶𝗰 𝗦𝗗𝗟𝗖 / 𝗔𝗜-𝗗𝗟𝗖 • The advantage isn’t just access to a model. It’s 𝘄𝗼𝗿𝗸𝗳𝗹𝗼𝘄 𝗱𝗲𝘀𝗶𝗴𝗻, 𝘀𝗵𝗮𝗿𝗲𝗱 𝗰𝗼𝗻𝘁𝗲𝘅𝘁, 𝗴𝘂𝗮𝗿𝗱𝗿𝗮𝗶𝗹𝘀, 𝗮𝗻𝗱 𝗳𝗲𝗲𝗱𝗯𝗮𝗰𝗸 𝗹𝗼𝗼𝗽𝘀. The shift is from using AI as a separate assistant to embedding it into the delivery lifecycle. • 𝗧𝗵𝗶𝗻𝗸 𝗯𝗲𝘆𝗼𝗻𝗱 𝗰𝗼𝗱𝗲 𝗴𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝗼𝗻. Claude (or any Model) can plan, draft PRs, review changes, summarize findings, and support cross-functional workflows. • Make plans, rules, skills, and 𝗮𝗴𝗲𝗻𝘁 𝗱𝗲𝗳𝗶𝗻𝗶𝘁𝗶𝗼𝗻𝘀 𝘃𝗲𝗿𝘀𝗶𝗼𝗻-𝗰𝗼𝗻𝘁𝗿𝗼𝗹𝗹𝗲𝗱 𝗮𝗿𝘁𝗶𝗳𝗮𝗰𝘁𝘀 that teams can share and evolve through Git. Use plugins to distribute shared skills, connectors, and agents across teams. • Use 𝗹𝗮𝘆𝗲𝗿𝗲𝗱 𝗴𝗼𝘃𝗲𝗿𝗻𝗮𝗻𝗰𝗲: deterministic checks → agent review → human approval. Traditional scanners provide coverage; models add reasoning ; humans retain final authority. • Treat 𝗽𝗹𝗮𝗻𝗻𝗶𝗻𝗴 → 𝗲𝘅𝗲𝗰𝘂𝘁𝗶𝗼𝗻 𝗰𝗼𝗻𝘁𝗲𝘅𝘁 𝘁𝗿𝗮𝗻𝘀𝗳𝗲𝗿 as a first-class design problem. Preserving intent and decisions upstream directly impacts execution downstream. • As engineering accelerates, 𝘁𝗵𝗲 𝗯𝗼𝘁𝘁𝗹𝗲𝗻𝗲𝗰𝗸 𝗺𝗼𝘃𝗲𝘀 — to product, design, QA, GTM, or adoption. AI-DLC needs to address the full delivery system, not just developers. • 𝗘𝘃𝗮𝗹𝘀 𝗮𝗿𝗲 𝗶𝗺𝗽𝗼𝗿𝘁𝗮𝗻𝘁 — measure quality, monitor drift, capture expert feedback, and use that feedback to improve agents and workflows.

Apollo didn’t start by hiring 100 people in LATAM. They started with 5. That is the part most CEOs should pay attention to. Back in 2016, Apollo started working with CloudTask with five Customer Success reps in Latin America. They tested the talent. They tested the management model. They tested whether a distributed team could actually perform. Then they scaled. As Apollo grew, that team expanded into Customer Success and Account Executives. At one point, CloudTask had helped support 100+ people working with Apollo. This is what I call the barbell hiring model. On one side: A small number of exceptional executives, founders, product leaders, and GTM leaders close to customers, capital, and strategy. On the other: Large teams of exceptional operators in markets like Latin America executing repeatable work across Customer Success, Account Management, Engineering, Sales, Support, and Operations. And now there is a third layer: AI across the entire company. ChatGPT. Claude. Gemini. Agents. Automation. You don't need 100 people sitting in an expensive office to build a serious company anymore. You might need 5 extraordinary leaders in San Francisco or New York. Then 25, 50, or 100 great people distributed across LATAM. And AI making every one of them more productive. The key is NOT moving 100 jobs overnight. Apollo showed the smarter path: 5 → prove it 10 → refine it 25 → operationalize it 100 → scale it That is very different from “outsourcing.” It is workforce architecture. Direct-hire demand creation. Manage repeatable execution in LATAM. Give everyone AI. The question I think every CEO should ask: Of the next 10 people you plan to hire, how many genuinely need to live near headquarters? Full Apollo story https://lnkd.in/eUg3ZUB8

The modern GTM stack is not Clay vs n8n vs Claude Code. The more useful setup is giving each one a job. Clay is still good when I want to see the data, inspect the list and iterate quickly. Automation tools connect steps. Code agents can handle custom scrapers, API workflows and enrichment waterfalls. One operator in the 2026 material described building a scraper with Claude Code, exposing it through an API, then plugging that API back into Clay. That is the direction I find more interesting. Not "which tool wins?" How do the tools work together?

The modern GTM stack is not Clay vs n8n vs Claude Code. The more useful setup is giving each one a job. Clay is still good when I want to see the data, inspect the list and iterate quickly. Automation tools connect steps. Code agents can handle custom scrapers, API workflows and enrichment waterfalls. One operator in the 2026 material described building a scraper with Claude Code, exposing it through an API, then plugging that API back into Clay. That is the direction I find more interesting. Not "which tool wins?" How do the tools work together?

Your GTM workflows should be in github right now. maybe not a hot take, but i just got off a call where everything was either in spreadsheets or siloed in 6 different tools. makes no sense. grab your API keys. build a knowledge base with claude, codex, whatever agent you use. version control it. then you just ask: what’s my deal state? what were the reply rates on the last campaign? no digging through 500 tabs. one answer, with evidence, that tells you how to make the next campaign better. Drake gets it 🤣

If you’re spending hours in Claude building lists of potential buyers, researching accounts, or identifying expansion opportunities, we can help you turn that time into action. Our new course inside ZoomInfo University, Getting Started with ZoomInfo MCP for Claude Users, shows you how to connect Claude to ZI MCP and access real-time B2B data. By the end of the course, you’ll have a Claude workspace tailored to your GTM needs— so you can find the right information in seconds instead of hours, without leaving the LLM you already use. 🤖Enroll today: https://okt.to/4QzShZ

AI’s in the news this week. (Shocker, right?) I’m going to skip right over Bill Gates and taxing bots because I think there’s probably been enough commentary on that. But one thing that I think is really worth paying attention to is this article from TechCrunch on three different AI deals that could affect how you’re planning roadmap and building product. NVIDIA is of course mentioned in relation to a reported deal to acquire Hugging Face for $13B which comes on the heels of a $6B deal to license Poolside's model-training technology and to bring on over 100 of its staff. This doesn't technically count as an acquisition as Poolside stays independent and is in fact raising its own $1B round at the moment. Stripe has also reportedly agreed to acquire OpenRouter for $7B+ which gives them control over the gateway startup that routes developer requests across different LLMs. That’s a lot of Benjamins changing hands. But what does it mean? - Choosing an AI platform to build on isn’t a one and done exercise. - Specifically with the Poolside deal, NVIDIA is showing once again that it's being careful to make distinctions between acquisition and licensing which is the difference between owning a company and just owning the upside. Since this is the third deal in nine months handled this way, it's unlikely to be the last. - Stripe's acquisition is interesting in its own right as it shows their bet is not on one specific model over another but infrastructure around deciding which model gets used and how that's billed. - For coding and building agents or performing other tasks that require back and forth communication with the LLM, frontier models like Anthropic’s Claude or OpenAI’s ChatGPT are still the better choice. If you’re building your product on top of someone else’s architecture, you and your team should understand what risk these acquisitions pose. How dependent is your roadmap on that of the AI tool? Could you migrate quickly if you had to? What would the implications of that be? Link to the full article: https://lnkd.in/gBHWAPrP #ArtificialIntelligence #B2BSaaS #GTM #StartupFounders

Just got off a session with Kushagra T. on GTM Engineering and this was honestly such a good conversation. We started with Clay and outbound, and somehow ended up talking about Claude Code, databases, APIs, AI agents, workflows, CRM and how all of this fits together. One thing I really liked was the idea of actually understanding the systems you’re building. If you’re using Claude Code, you don’t necessarily need to know how to code everything yourself. But you should be able to look at what it built and understand what’s happening. We also got into the whole AI agents conversation. What actually needs to be an agent? What can just be a workflow? That part was really interesting. The coding agent discussion was probably my favourite. Instead of throwing a huge CSV at Claude Code and saying “research all of this”, you can structure the work using scripts, APIs and different models for different parts of the process. You basically give the agent a job to orchestrate rather than asking it to do everything. That was a really useful way for me to think about it. Really enjoyed this one. 😊 Thanks Kushagra T. for taking the time to do this with us, and Yogesh Jaiswal for putting the whole session together. 🙌

Most AI in your GTM stack has amnesia. It has no memory, reaches the same conclusion it reached last week, and forgets it the second the table refreshes. And you pay for that lookup every single time. Clay's new Account Agents change that. Think of them as the stateful cousin of a Claygent. Instead of firing once per cell in a table, they run per account across a whole segment, and they remember what they concluded last time. If a signal lands, the following happens: The agent shows up already knowing the account: past touches, deals, calls, prior conclusions. It reasons over what changed, then picks a next step from a list of actions you pre-approved. The conclusion and the reasoning get written back onto the account, so the next run starts from there. Then Clay executes: update the CRM, notify the owner, run the next step of the play. A normal Claygent starts from zero every time. An Account Agent builds intelligence account by account. That's the difference between "re-research this company for the fourth time" and "here's what changed since last week." Where I am using it: Expansion, cross-sell, and upsell. Inbound routing when you had multiple form fills. Closed-lost revival. Is your stack still re-researching the same accounts every week?

Building a GTM stack in 2026? Here's one tool per layer, nothing extra. You need 9 that actually work together. Here's the trimmed stack for a lean GTM engine in 2026: 1️⃣ Finding Prospects (Database) • Airscale, Find and enrich your ICP easily • LinkedIn Sales Nav, The OG of lead finders 2️⃣ Finding & Verifying Emails • QuickEnrich, B2B email finder and verifier API • ZeroBounce, Cleans and validates your lists 3️⃣ Finding Phone Numbers • Airscale, Accurate, scalable waterfall enrichment • Kaspr, Easy LinkedIn phone extraction 4️⃣ Cold Email Outreach • Instantly.ai, High deliverability and automations 5️⃣ LinkedIn DMs • HeyReach.io, Automated LinkedIn messaging 6️⃣ Multichannel Outreach • lemlist, Personalized multi-touch campaigns 7️⃣ Web Scraping & Enrichment • Claygent, AI-powered data extraction • PhantomBuster, Versatile scraping and automation 8️⃣ Infra and Deliverability • Zapmail, Reliable Google and Outlook mailboxes 9️⃣ Intent Data • Prontohq, Signals from job changes and new hires Nine tools. Every layer covered. No overlap slowing you down. Save this and revisit when you're plugging a gap in your outreach.

Stop Guessing. Start Selling with GTM 🦋 Signals. 🎯 Your next high-value lead may already be showing buying signals. 👀 Track job changes, funding rounds & new job postings automatically. Then use Claygent to uncover custom signals from Google, call transcripts & product data. 🤖 The result? Know who to target, why now, and when to reach out. 🚀 #Clay #LeadGeneration #SalesIntelligence #RevOps #B2BSales

📍 I read this 50 times a day because I am just chronically online on LinkedIn. Claude Code replaces Clay. I mean… => If you don't have the budget for Clay, Claude Code is gold. But there are nuances. => Claude Code only replaces Clay if you can build what Clay already built. A crawler that goes page by page, scrapes it, and actually understands what it found. (P.S. I am a fan girl of Claygent). But if you can not do that, you just have a dumb chat window or a bot (or an Agent in fancy terms). The problem starts when you start to push the most basic google sheet text manipulation to Claude code. Google sheet is free but tokens will keep getting pricier. It becomes inefficient. Clay does these manipulations free of cost if you do not understand formulas in Google Sheets. Clay isn't priced for small agencies. It's priced for a team of 60 running 100 clients a month. And at that scale it isn't even a decision. Then the right question is how efficiently you can get the outcome with respect to your team size. Do you agree if you are leading a team of 10 people?

I found a way to see everyone who engages with my competitors' posts. here is the stack I used, in the order I ran it. I wanted a warmer list than a cold export. people commenting under posts in my category already care about the topic, and LinkedIn keeps those names public. so I wired six tools together and worked back from one post. - I start with phantombuster's activity extractor, pointed at a competitor's profile, to pull their recent posts with like and comment counts. that told me which post actually landed. it does not export who engaged, so that came next: https://lnkd.in/geMkAiDd - then the post commenters export. I run this one before the likers list, because someone who wrote a sentence is warmer than someone who tapped a button, and I get their actual words to open with. I kept it under 50 posts a day: https://lnkd.in/gkevdHkd - the post likers export widens it out. LinkedIn only displays 3,000 likers on a post, so that capped how far I could take any single one: https://lnkd.in/gtqZhBmb - clay's waterfall finds the work emails. it tries over 150 databases one after another until one comes back with a match, which stopped me losing rows when a single provider had nothing: https://lnkd.in/gC6eCRgU - claygent sits in that same clay table. I feed it the person's role and their comment, then ask it to score fit from 1 to 10 and draft one opening line quoting what they wrote: https://lnkd.in/giNDaUqE - instantly sends the sequence. every plan includes unlimited sending accounts, so I spread the volume and kept each domain under its daily limit: https://lnkd.in/gNRyyjbj I have run this against a handful of competitor posts now. the commenters list has out-converted the likers list every single time. happy to walk through how the clay table is wired if that is the part you want, just say so in the comments.

We found a way to see everyone who engages with a competitor's posts. Here is the stack we used, in the order we ran it. We wanted a warmer list than a cold export. People commenting under posts in a category already care about the topic, and LinkedIn keeps those names public. So we wired six tools together and worked back from one post. - We start with PhantomBuster's Activity Extractor, pointed at a competitor's profile, to pull their recent posts with like and comment counts. That showed us which post actually landed. It does not export who engaged, so that came next: https://lnkd.in/gGi637Nd - Then the Post Commenters Export. We run this one before the likers list, because someone who wrote a sentence is warmer than someone who tapped a button, and we get their actual words to open with. We kept it under 50 posts a day: https://lnkd.in/gcn-CNun - The Post Likers Export widens it out. LinkedIn only displays 3,000 likers on a post, so that capped how far we could take any single one: https://lnkd.in/gES9jRm8 - Clay's Waterfall finds the work emails. It tries over 150 databases one after another until one comes back with a match, which stopped us losing rows when a single provider had nothing: https://lnkd.in/gUNc8UEi - Claygent sits in that same Clay table. We feed it the person's role and their comment, then ask it to score fit from 1 to 10 and draft one opening line quoting what they wrote: https://lnkd.in/gfv-xs52 - Instantly sends the sequence. Every plan includes unlimited sending accounts, so we spread the volume and kept each domain under its daily limit: https://lnkd.in/g__x57VS At Supernodes we have run this against a handful of competitor posts now. The commenters list has out-converted the likers list every single time. Happy to walk through how the Clay table is wired if that is the part you want, just say so in the comments.

Last week I posted about how proud I was of my "wonderful" first Clay build. Multiple tabs, interdependencies, waterfalls, Claygent formulas, run conditions…all of it stacked on top of each other. BUT… Then I ran it against a small batch of low-priority contacts, and it was a MESS. I did what every first-time builder does: I overbuilt before I validated anything. Then, I tore it apart and rebuilt it by hand, with far less help from Claude the second time around. It was BRUTAL…but the 2nd time was when I really learned Clay. There are lessons buried in Techstars' book "Do More Faster" that I read years ago that I still think about: don't overbuild, start minimal and fail fast. Brother Tom Fahey at Saint John's High School used to say the same thing in four words: "Keep it simple, stupid (KISS)!" One step at a time... I learned a version of this in grad school too… A professor explaining business expansion put it this way: Change ONE dimension at a time. Either: A) New product OR B) Different geography. Never both at once, because you can't tell which change actually worked or broke. The dopamine hit from adding more: more enrichment, more sophistication, more scope, isn't the same as building something that actually works. Validate one thing. Keep it boring, one step at a time, and validate. Then earn the right to add the next one. Can you think of any other domains in life where it’s hard to restrain yourself from “overbuilding” before you’ve earned the right?

“CoLd eMaiL tO E-cOmMeRce is DeAd!” ok. here’s how we generated 45 leads in 20 days for this ecom agency: 1. Create a STRONG frontend offer (free asset/work) Everyone is spamming e-com with the same crappy pitches. “We guarantee to increase email revenue by 50% in 90 days or u don’t pay!” “We’ll create a landing page that increases conversions by 20%!” “We’ll 2X your ROAS or I’ll buy you a steak dinner!” This slop has been pushed since 2022, THIS outreach is dead. What’s not dead is actually offering something people want, regardless how saturated the industry is. What that looks like: - Free case study masked as a “playbook” - Free sample of your product/service - Free micro service that isn’t your main service (just a lead magnet) These work, and brands still want them when positioned right. 2. Hyper-qualify lead lists E-commerce data is typically pretty bad for some reason across most databases. Even when you filter for 11+ employees, you’ll run into brands that somehow have less than a $1k/mo budget for marketing spend lol. So you need to qualify further than the base filters: - Run claygent to confirm is an Ecom brand - Score the brands online presence (signal whether have budget or not) - Filter for E-com technologies (Shopify, WooCommerce, etc) - Verify whether they’re actively running Meta , Google, LinkedIn, Apple Ads or not with Apify Apollo.io alone is not going to give you a well targeted list, need to qualify further. 3. Very short & casual messaging Compared to most industries, E-commerce is probably the most “laid back” when it comes to messaging formalities. From what we’ve seen, the more formal you are the worse emails perform. As well as email length – these guys are receiving 20-100 cold emails/day, they need to be able to understand what your entire email is about in <5 seconds or they’ll just skip it. The best performing email we’ve ever sent to E-com brands was a one-line email that is less than 20 words. It’s along the lines of “John, interested in XYZ thing I created for your brand?” Print leads and meetings, which really came down to the offer. Message just presented the offer in a quick and simple way. — Email to ecommerce is really not the beast a lot of people in the space make it out to be, there’s just less margin for error. Follow these frameworks and you’ll be golden.

I counted the platforms we touch to get one GTM outcome. Five. Claude, re-prompted because the answer wasn't quite right. Cursor, writing into repos nobody has fully mapped. Zapier, holding a workflow together that one person understands. Claygent, enriching records nobody has audited. LangChain, running agents that don't talk to each other. Each one is good at its own job. The problem lives between them. None of them share memory, so context you established in one is invisible to the next and you rebuild it by hand every time. That rebuild is unpaid work nobody puts on an invoice. Add up the subscriptions. Then add the hours spent moving context between them. That second number is the real invoice.

At our recent NYC meetup with Clay, Head of AI Jeff Barg, ML Engineer Vyshnavi Khota and Software Engineer Soroush Khadem shared how they scaled agent evals to 300 million runs a month. Watch the full talk to learn: ✅ How Claygent and Sculptor got to production scale ✅ Clay's four quadrant framework for agentic evals ✅ Why closing the loop between production and offline evals is the hardest part ✅ How a data lake and long context are changing what agents can do with data Check it out on YouTube: https://lnkd.in/grStyZ4m

That's a wrap on my summer at SeaCube Container Leasing! These past 13 weeks, I had the opportunity to work as an Ai and Security intern and I couldn't have asked for a more hands-on experience. During my internship, I was involved in 2 major projects, network drive migration and Clay go-to-market ai integration. By far, my biggest project was the network drive migration project which impacted over 150 users across 4 international offices. The objective of this project was to move away from Administrative Templates, which was the original Microsoft Intune policy that mapped the drives, however that was being deprecated. A more modern solution to this was Scripts and Remediations through Intune which allowed PowerShell scripts to run hourly. The first script would detect and if detection failed, remediation would run. This structure followed a drive + breadcrumb logic which made self healing much easier incase a network drive drops. Beyond the infrastructure side, I gained a lot of useful ai + marketing knowledge through my project on Clay GTM, where I built an AI-enriched outbound sales pipeline. Using custom Claygents (clay ai agents), news RSS feeds, and website scraping, I helped identify and enrich quality leads worth pursing for cold storage container rentals. Overall, I delivered 4 lead enriched workbooks which resulted in around 400+ total leads across major US shipping regions! I want to take a moment to thank a few people for making my internship so memorable and meaningful! First my manager Sal Narciso for giving me this wonderful opportunity to grow in the field of IT and expand my knowledge into AI! Next I'd like to thank my fellow IT members Michael P., Sonali Rana, Sky Arizankov, and Niken Patel for aiding me in my network drive migration project, without them, I would not be able to test and deploy such an important change in the company infrastructure! Lastly, I would like to thank my coworkers over at SeaCube Cold Solutions, Patrick Cooper and Adam Janvey for giving me the chance to incorporate Clay GTM into the company! Their constant support and assistance helped me produce such a robust ai-enriched lead generating pipeline! Once again, huge thank you to the entire SeaCube Container Leasing team for giving me the opportunity with real, impactful projects and for all the mentorship along the way. Excited to carry everything I learned into my next chapter at NJIT! #internship #InformationTechnology #NetworkSecurity #AI

Deepline is hiring a Founding Head of Growth! I've gotten to work directly with both Jai Toor and Chirag Toprani on a number of marketing projects and can personally vouch that this is an awesome role. Do I know anyone looking to take on a new adventure? Link in comments.

This is great, and it's more complicated: A couple recent conversations with GTM Engineers & Applied AI people in the revenue space had me thinking there's 1 big shift happening, This matters for anyone trying to advance their career in GTM or get into GTM in the first place Early 2026: Claude code was popularized, suddenly a decent chunk of Clay's value disappeared, people realized you can build lists with claude + a couple data providers, prospeo.io / FullEnrich etc. people started building, Fable 5 was released, GPT 5.6 sol right after, suddenly you could source your whole market into a database and use APIs to pull anything you might need. It's now August of 2026 and companies have already been hiring for GTM for a while, yet the role is already changing Claude code made execution even cheaper, companies like Deepline made data cheap, so now anyone can build. But if you can build anything, how do you know what's worth building? A lot of GTM talent is obsessed with the next workflow, the next dashboard, etc. They're told the VP sales needs help with closed-won analysis, so they start building right away but did they ever sit down to listen to the VP sales, and understand what they want? Right now I see only 2 ways a GTME could realistically advance their career: 1. Get really good at understanding the problem, talking with the CSM, the head of sales, the account executives and: 2. Creating a specific scope of *what* needs to be built, and what the end goal is. If we look at historical data of AI advancement, agents will be able to do 95% of what an "executioner" GTM engineer does right now, building lists, workflows, dashboards, etc. So where the value lies is in understanding their problem, and creating a plan of what needs to be built, the rest will be done by an agent soon, not a junior GTME

हर खामोशी अहंकार नहीं होती… कुछ लोग अपने हालात से लड़ रहे होते हैं। ❤️‍🩹 #DeepLines #LifeReality #SilentBattles #Emotions #LifeLessons #StayKind #Understanding #TruthOfLife #HeartTouching #PositiveThoughts

I have built company research and scraping workflows, cleaned and deduplicated data, worked with Clay for enrichment and signals, explored job-change and ICP qualification, and started turning those pieces into repeatable pipelines. But doing this manually also showed me where the real gaps are. A lot of the work involves stitching providers together, moving data between tools, testing enrichment sources, building waterfalls, validating outputs, and figuring out what should happen next. That’s why Deepline caught my attention. Deepline is building the infrastructure that lets GTM workflows run through agents, connect data providers, enrich and validate records, and turn repeatable GTM processes into systems rather than endless manual work. Many of the things I have already done could have been faster, more scalable, and more repeatable with the right infrastructure behind them. Winning the Bootprint 🥾 (prev Clay Bootcamp) 🥾 GTM Engineering Scholarship would give me the opportunity to go deeper not just learning tools, but learning how to design better GTM systems, automate intelligently, test faster, and build workflows that can actually scale. And to Deepline, thank you for sponsoring the scholarship and investing in people who are learning to build this future of GTM. Thanks to Nathan Lippi 🥾Bootprint 🥾 (prev Clay Bootcamp) RevenueHoop FastForward EmailBison OutboundSync Clay Girls Who Clay community I am excited about what comes next. 🚀

I'm super excited to announce the first official lineup of events at the AI Campus Barcelona. We are finally opening in September! Some of our featured sessions: - Holded is sharing how AI has transformed their Customer Support operation, firsthand from their Head of Customer Support Javier Gonzalez Valenzuela - Enginy is hosting an expert panel on the new rules in B2B sales for SMBs, moderated by their founder and CEO Kai Brandt - Women Techmakers Barcelona are hosting a session on how to become an AI-native Product Manager with Mireia Sangalo from Vonage And there's more: - Agentic prospecting with Claude Code + Deepline by Oriol Serra from the Conversion Architects - Platform Selling: Positioning Value in the AI Era by Zach Gropper from Insight Revenue - AI-Augmented Product Management by Gabriella Martins from Docplanner - How to make your pre-AI pricing future-proof by Krzysztof Szyszkiewicz from Valueships - Design your own perfume with AI by Enrico Gavagnin and Lucas Genton from Visium - AI Adoption: Before your Team Buys More AI Tools by Lucré van der Zwaag and Nick van Zuijlen from the Bold Minds Collective Registrations are open so claim your spot now, while half the city is still on their August break. You can find the full program on our website and in the comments. Hope to see you at one of our sessions!

Mapping the 19 sales tools I'm watching in 2026. City by city, west to east. A few are here because I know the founding teams personally and believe in what they're building. A few have been around long enough that I'm confident they're not going anywhere. And a few just made the cut because I genuinely love using them. Bay Area: Apollo.io, Caretta, Ordinal, Cargo 🧱, FullEnrich Seattle: Outreach and ZoomInfo Atlanta: Salesloft Toronto: Prospeo New York: Clay, Deepline Boston: HubSpot Paris: lemlist Lyon: Breakcold Zürich: OXYGEN & TRICLE AI North Macedonia: HeyReach Tallinn: Instantly.ai Sydney: Smartlead