If I wanted to Become a GTM Engineer… I'd do it in this order. Earlier this week, a student emailed me with a simple question. "If you were starting from scratch in 2026, would you learn Clay first or Claude Code?" I've been getting variations of this question a lot recently. Usually followed by: "How do I land one of those $5k to $7k/month remote GTM Engineering jobs?" My answer usually surprises people. In this week's Clay Cult Newsletter, I break down the roadmap I'd follow if I had to start all over again. Including: → Why learning B2B sales is more valuable than learning another tool → When Claude Code should become part of your stack → Why Clay makes much more sense once you understand the workflow → The mistake I see almost every beginner make → The simple 5 question checklist I use before building any automation If you're trying to break into GTM Engineering, or you're wondering what to learn next, this issue should save you months of heading in the wrong direction. You can read the full newsletter here: https://lnkd.in/e4avKX9b

2026 AI FIRST GTM Tech Stack A lot has changed since last year (crazy how fast) so here's a breakdown of the most interesting things I discovered, tested and put in my s++ TIER LIST General Productivity Wispr Flow Claude Design GTM Ops Claude + Deepline (93 integrations in 1) Clay (best for enterprise) Email Sending Tools EmailBison Smartlead Instantly.ai LinkedIn Tools Hey Reach (best for linkedin automation) Sales Navigator (best for network and social selling) Cold Calling Dialers Salesfinity Nooks Intent Signal Providers HG Insights (most accurate tech stack data for enterprises) Vitelis (most powerful open data engine right now) RB2B (company visitors) Mobile Enrichment Forager (best mobile coverage) FullEnrich (best waterfall of providers) Cold Email Infrastructure sending.ac (Microsoft Azure domains + addresses) 1-1 Outbound Tools Outreach (best for enterprise and big team) Commsor 🦕 (warm intro and GTN) CRM Salesforce (best for enterprises) HubSpot (best for scale-ups) Attio (best for brokies) What's the single most interesting piece of tech stack you added in 2026? - Do you want my best AI & GTM Engineering Trainings? Go to Michael Saruggia and click on 'View Newsletter' | 24 comments on LinkedIn

3 boring ways to use AI that can get you high intent leads every day automatically. Multiple tools can set this up for you, you just need to pick your favorite. Pick your favorite AI coding harness. Cursor, Codex, Claude code. Anything that can run a scheduled task every day. These tools can do all of these strategies on their own. But consider layering Parallel Web Systems or Exa on top for some more purpose built searching. However you want to do it, just ask your AI buddy to do any of these three things. I want to be clear that all of these start with defining who your very best customers are. 1. Ask AI to find 3 perfect companies based on some trigger event per day. Keep a local csv so it doesn't show you the same 3 companies again for whatever time duration you want to set it for. I had Parallel find 3 companies that are hiring for GTM Engineer roles as you can see in the image. It's boring but getting this data collection done on a daily basis with something as smart as the parallel.ai task agent is a great use case. 2. Ask AI to find an industry event that it is coming up and find the sponsors. If we could run only event campaigns on cold email for the rest of time, we would be the most successful agency ever. Of course, we can't because there aren't a ton of events. But when they do come, you want an Agent finding the new ones and scraping the sponsors / attendees if possible because the list is so high intent. 3. Automate your Dream 100 outreach When people ask me how outbound is going to have to be different, I consistently believe that it will evolve to have to bring true value in the message. Chet Holmes created the dream 100 outreach as a thought process of "what are the lengths you would go to acquire your dream 100 customers." Once you answer that question, most of the things you would do are totally possible for AI to do on your behalf. You can use your token allotment to run a skill on 5-10 companies per day or more and create the dream 100 resource and then you can manually send (or automatically of course) to those prospects. My second post today will go over one we made yesterday. Copy and paste this to your agent friends and comment how easy it is to set up. It's something I think anyone can do no matter your AI skill level.

80% of outbound advice fails for one reason. It assumes Cold Email, LinkedIn, and Cold Calling are interchangeable. The cold email people think everything can be solved with more volume. The LinkedIn people think content fixes everything. The cold calling people think everyone should just pick up the phone. Reality? They're all trying to solve different problems. – Cold email is for scale. – LinkedIn is for trust. – Cold calling is for feedback. That's why the best outbound teams don't choose one. They use all three. Different channels. Different jobs. One pipeline. Which one are you relying on most right now? 👇 | 85 comments on LinkedIn

Pick the channel based on the account value, not the other way around. Here is the rule we use across scaling B2B SaaS: What each channel does: ↳ Cold email = volume engine. 500 to 2,000 sends a day. Direct pitch is fine. ↳ LinkedIn DMs = precision play. 40 DMs a day, capped. Value first, no hard ask. ↳ Cold calling = breakthrough channel. 50 to 100 dials a day. Real-time, adaptive. ↳ Paid ads = warming layer. Familiarity before any outreach lands. The rule we follow: ↳ Tier 1 accounts get all 4 channels orchestrated together. Paid ads warm. LinkedIn nurtures. Cold call breaks through. Cold email only if the offer fits. ↳ Tier 2: get cold email at scale, plus paid ads as the warming layer. ↳ Tier 3: No LinkedIn DM. No cold call, just emails at scale. What breaks the rule: ↳ Push Tier 1 into a cold email campaign, and you burn the relationship ↳ Scale LinkedIn past 40 DMs a day, and your LinkedIn stops working (you can’t reach all leads if LinkedIn connection requests are capped at 100 / week) ↳ Run cold calls with no warming layer, and connect rates are lower ↳ Run paid ads with no outbound Tools to use: Email: Instantly, Smartlead, EmailBison + EmailGuard, MailReach (deliverability) Ads: ZenABM , Fibbler, Satlo, FactorsAI (mid-market), 6sense, Demandbase (enterprise) LinkedIn: Heyreach, lemlist , La GrowthMachine Prioritize the best accounts and contacts with all touchpoints your current GTM motion can do. Decrease the efforts that go into low-fit accounts. ROI = maxed + your SDR team won’t be drowning in a list of activities and reports to complete Full breakdown in the chart. P.S. Are your sales reps spending too much time on the wrong accounts, contacts, or drowning in to-dos because they don’t know where to prioritize? Book a time with us, and we can help orchestrate and connect all 1st, 2nd, and 3rd-party signals into 1 allbound motion.

Nobody in cold email would dare put a link in the first message. Even though it would reduce friction in the booking process, everyone knows it ruins deliverability at scale. So we've become so accustomed to not writing copy with links. But LinkedIn InMails are different in that those deliverability issues don’t exist. So, I do not understand why more of us are not taking advantage of this and pairing a very strong offer with a link to book a call (I say this as we are now ramping this up heavily). If you have experience with this, drop it in the comments. Would love to see. | 25 comments on LinkedIn

I’ve come to realize that 50% of the companies I talk to use Clay solely as a data tool. It’s great at that, but there is SO much more you can use it for. Here are some other use cases to consider: 1. ABM Asset Generation Do you or your team spend way too much time creating slide decks, pdfs, case studies, and other marketing assets to send to prospects? You’re not alone, and this is a solvable problem. Claude has paved the way for generating high-quality marketing assets, but it has a harder time doing this at a high scale- think hundreds of accounts. Clay covers this use case well; you can even create an external agent that generates the assets, connect it via API to Clay, and take advantage of their automation infrastructure to do this for all of your accounts at once. 2. Deep Research Briefs Finding prospect data is one thing, creating a brief that your sales team actually reads is another. So many research automations fail because the data is just barfed out to reps without any regard for how they prefer to ingest it. Clay can both find that data and then use frontier models to sequence it into whatever format your reps deem most helpful. 3. Automated Outbound (and not just fully automated) A huge misconception from many Clay-curious people is that their outbound automation capabilities are only good for the spray and pray stuff. This couldn’t be further from the truth; tons of our enterprise accounts use Clay for man-in-the-loop automation as well. I see outbound automation as a spectrum. On one side, there’s a system where only the most tedious of tasks (like lead routing) is automated. On the other side, everything is done without rep intervention at all. The reality is, 80% of the businesses we work with fall in the middle. Don’t throw Clay out just because you see an outbound agency using it to send 1M emails a month. This is just scratching the surface, but hopefully is helpful and sparks some ideas on how it could be used! | 17 comments on LinkedIn

I ranked Sales Signals on a tier list to invite the LinkedIn world to disagree with me. Did I get anything wrong/would you change any of the rankings? Explanation below. I think people talk more about signals than ever before (myself included). They are sexy and just feel like the right thing to do. Often times, more effort is put into the signals than your offer which is a mistake. We just made a change to our cold traffic offer that is converting better with 0 signals than the high signal campaigns with a different offer. Anyway. Here's a list of common sales signals and how I'd rank their importance. Disclaimer: I'm viewing this from the lens of how effective is the signal if I were giving advice to a room of 200 people never doing cold email. I rank M&A activity as D tier but that might be the only thing you care about. Take this as general advice. Here's the list of signals I rank in the article. Custom Trigger New in Role Lookalike Campaigns Other General Growth Signals Competitor Movement Customer Complaints Job Post Data Tech Stack Signals Website Visitors Social Activity Department Growth Lawsuits / Compliance Issues Funding Pricing Change M&A / PE Activity Cost Cutting / Layoffs Product Launch Let me know if you have a different opinion on any of them as broad market advice. I completely agree that I am wrong when giving this general advice for specific circumstances that these signals crush for. | 23 comments on LinkedIn

Why great GTM engineers must be terrible copywriters (at least at first) Some of my best cold email copies have often come from a feedback loop. Here's how: --- Step 1: We build the foundation Client sends over their onboarding doc, ICP, existing copy, case studies, anything that works in their niche. I read all of it. Then I read it again. If the client has learnings from past campaigns, I take those seriously. What they tried. What worked. What didn't. --- Step 2: AI generates the first batch I run everything through Gemini. Claude Code handles the research layer. First, we go wide. 5 strategies minimum. For bigger clients, we launch with 10 to 20 angles right off the bat. Long copy. Short copy. Evidence-based. Self-deprecating. Whatever the offer calls for. --- Step 3: We go back to the client I don't hide behind the AI. I tell them exactly what we tested, why we chose these angles, and what we're expecting to learn. Clients appreciate the honesty. Because they are paying for me to run the right experiments and read the results. My goal is to engineer Message Market Fit as quickly as possible for my clients. --- Step 4: Feedback lands. I believe most people drop the ball at this stage. Feedback context is a gold mine. If a client tells me "this angle flopped" but can't explain why, I dig deeper. Because sometimes the problem is bad targeting. --- Step 5: We iterate We take the feedback, adjust, and run it again. Week after week, the copy gets sharper. Because the audience is telling us what they want. We just have to listen. --- And that is how we engineer Message Market Fit for out clients in less than 90 days. P.S. We're offering a free trial: 6k leads contacted, 3 unique email campaigns, turned around in 48hrs. Apply here https://lnkd.in/eCcbZDRE

We send 2 million cold emails a month and generate 3,000+ positive replies. And our pre-send checklist has only 4 steps (CRITICAL ONES) You DON'T need 15-20 steps checklist. A checklist only works if you actually run it every single time. Here are the 4 things we check before any campaign goes live. 1. Copy approved by the client ✅ Get the client to read it first. Because you can recover from a bad open rate. But you can't recover a client who doesn't trust you. 2. Targeting verified 3. All emails verified ✅ Stack multiple providers - BounceBan, Reoon, Icypeas, Prospeo. 4. Send timing set to the recipient's timezone ✅ Send during prospects local timezone. Ex: If you're in Europe sending to US prospects at 9am your time, that email is landing at 3am their time. Which will get buried in their inbox before they even wake up. That's it. Every campaign goes through those steps. What's on your pre-send checklist?

Browser Use open sourced another amazing tool for us in GTM they are calling browser fetch. A tool to turn any website page into markdown for free. Here's where to get it. Browser use has been pumping out some of my favorite open source and free tech. - Browser harness for phenomenal browser automation - we're using it build Clay tables for free. - Video use - helping me edit videos and post to a secret second channel to see some organic experiments we are trying to learn from. And now brwoser fetch - a free way for us to scrape web content from pages Markdown or JSON, ready to use from any URL. > renders JS & beats Cloudflare, bot walls, and IP bans > built-in proxies — no scrapers, no per-site APIs to maintain > fan out across thousands of pages at once, zero rate limits Huge to add to your Claude Code/Codex Md files right now. | 19 comments on LinkedIn

If you want to AI-proof your job in outbound, you have to do everything you can to move towards the one part AI will never replace: The actual closing of the deal. AI is currently eating away at the lower-level SDR tasks (qualification, call booking, etc.) I don’t think it’ll ever replace an enterprise sales rep. In my view, here’s a ranking of the parts of the outbound / SDR most exposed to AI replacement vs. least: 1. Inbound SDR / inbox management 2. Lead gen and list building 3. Outbound SDR work 4. SDR team management 5. Sales and closing If you work in spots 1-2, your career will likely benefit from working towards 4-5. | 23 comments on LinkedIn

My LinkedIn content generated ORGANIC 500,000+ impressions in the last 90 days. Not because I post more. Because I built a content engine. Every Friday at 6 a.m., our best‑performing posts are automatically researched and drafted. I only fine‑tune them. Here is the exact system behind it: Step 1: Foundation (built once) ↳ Voice profile per person: Each team member has a defined writing voice and role ↳ ICP map: We map buyer tiers and the GTM problems they actually care about ↳ Content pillars: 3 repeatable themes tied to the pipeline ↳ Hook + CTA frameworks: Best‑performing structures saved as Claude skills ↳ Design system: Brand colors, layouts, and templates locked into Claude Tools: Claude Python Step 2: Research layer (runs weekly) ↳ Apify scrapes LinkedIn to detect high‑performing posts in our niche ↳ Fellow captures transcripts from client calls ↳ Internal Slack debates often become the best post ideas. Client conversations are our highest‑signal source. Every Friday at 6am, the system delivers a draft for the whole team. Tools: Apify Fellow - AI Meeting Assistant Claude Step 3: Production (90% AI, 10% human) ↳ 50+ hook variations generated and ranked by expected performance ↳ Claude drafts the post in each author’s voice ↳ Hook + body matching tests, which combination works best ↳ A 5-dimensional rubric grades tone, clarity, insight, and rewrite suggestions Tools: Claude Code Scripe Step 4: Repurpose the engine ↳ One high‑performing LinkedIn post = one validated idea ↳ The idea gets rebuilt for other formats ↳ Our “best‑performing database” updates daily This compounds the system. Step 5: Maintenance ↳ Python scripts maintain the automation via GitHub ↳ Monthly refresh of tone, data, and frameworks ↳ Weekly updates of best‑performing posts The longer this runs, the smarter the system gets. The real difference: Most teams use Claude like a better ChatGPT : Prompt. Copy. Paste. Forget. We built Claude Code as the operating system behind our content engine. That’s how one idea becomes a pipeline‑generating system. Have you started automating your content workflow yet? Comment ENGINE and I’ll send the full setup. P.S. If your team spends hours writing posts every week, you don’t need more writers. You need a content system. We can build that for you : - ) | 151 comments on LinkedIn

2024: “AI will replace SDRs, it’s cheaper & more effective” 2026: … API bills entered the chat They used to brag about cutting headcount with AI Just to realise that it might be cheaper to keep the human SDRs See, in 2026 people are waking up. What became clear: 1. Human in the loop will outperform AI only outreach 2. Good personalisation gets your replies, not meetings 3. AI’s good at personalisation, but doesn't get the context What does move the needle forward is a good offer. So, you need to A/B test at least 6 different offers to see what sticks 1 data 2 teardown 3 audit or diagnostic 4 checklist or benchmark 5 money back guarantee 6 freebie (Claude skill, Clay table, LGM cold email campaign setup, etc) Talking about A/B test, also don’t forget to test: - Subject lines - Openers - Social Proof - Call-to-action - Length of an email I always set up in LGM 6 different A/B tests going at the same time How do you A/B test stuff? | 158 comments on LinkedIn

>50% of outbound success is following up (tastefully) until you make the sale or get told to F off. This is a great example: 1. Person cold calls Daniel: he doesn't answer. 2. Tries a couple weeks later: he doesn't answer. 3. Tries 11 days later: They talk for 27 mins. Most sellers would have stopped after not getting a pick-up from call 1. The same goes for cold email. Outside of leads that explicitly say you're not interested, you should reach them again 30 days post-initial touch. Just because they aren't ready right now does not mean they won’t be eventually. | 41 comments on LinkedIn

>50% of outbound success is following up (tastefully) until you make the sale or get told to F off. This is a great example: 1. Person cold calls Daniel: he doesn't answer. 2. Tries a couple weeks later: he doesn't answer. 3. Tries 11 days later: They talk for 27 mins. Most sellers would have stopped after not getting a pick-up from call 1. The same goes for cold email. Outside of leads that explicitly say you're not interested, you should reach them again 30 days post-initial touch. Just because they aren't ready right now does not mean they won’t be eventually. | 39 comments on LinkedIn

I’ve bootstrapped $0-1m ARR 3 times, and I’m about to do a 4th time with MoltSets. Here’s my 5-step plan for creating a lean, mean, high-velocity growth machine: To set the stage: - Moltsets is unlimited contact data APIs built for Clay and Claude Code - It will be PAINFULLY simple, with a throw away price ($27/mo) - All customer inquiry (pre and post signup) will be handled by AI agents ——— Now, the 5-step playbook to get to $1m ARR for the 4th time. Step 1: Tap into the power of the “Founder Brand”. Thanks to Dave Gerhardt for putting me on to this. I’m going to create a huge amount of awareness for this by posting about it, doing a weekly live show, and making YT videos. Step 2: Ride an epic wave. The explosion of people posting about Claude on LinkedIn is second only to the explosion in Anthropic’s actual revenue in March ‘26. I want on the wave. Step 3: Give creators BIG visibility in exchange for UGC. I’m not exactly sure how I’m going to do this but I’m going to make a Claude-first weekly live workshop show. I believe my founder brand and that topic will pull a larger audience than creators can pull themselves by orders of magnitude. I will also engage with ANY post that tags MoltSets from my personal account. (DM me if you’re interested in being a creator.) Step 4: Keep the product painfully simple We are going to obsess over high quality data with disruptive subscription pricing. The product will only ever do 5 things - people search, company search, email enrichment, phone enrichment, and email validation. Step 5: Run it all w/ AI Simplicity will allow us to run it all customer interaction through Intercom Fin, like we did w/ RB2B. Robb Clarke will once again create world-class documentation that will make it all possible. TAKEAWAY We are still VERY early with MoltSets. There’s also a a few core things I have NO IDEA if they will work or not. But here is one thing I know FOR SURE that we’re doing right… 2025 was about building WITH AI. 2026 and beyond will be about building FOR AI. DON’T GET LEFT BEHIND Join the MoltSets waitlist here: https://lnkd.in/g8xXTCQr BETA starts soon! | 67 comments on LinkedIn

We've run enough campaigns to know exactly where outbound breaks. The data is always pointing at the same place. Most teams inherit an offer from their inbound messaging, paste it into a cold email, and wonder why nothing comes back. They blame the copy. They blame the channel. They test a new subject line. But when you look at what actually drives replies across every campaign we've run, the breakdown is always the same. 80% offer. 15% messaging. 5% infrastructure. Here's what each part actually means in practice: The Offer (80%) ↳ Cold audiences don't respond to repurposed inbound messaging. ↳ Before anything goes out, stress test it: outcome, mechanism, risk reversal, pilot path. Messaging (15%) ↳ If the offer is strong, copy just needs to communicate it clearly ↳ Rewriting copy on a weak offer is optimizing the wrong thing Infrastructure (5%) ↳ Deliverability, domain setup, and data quality decide if you reach the inbox ↳ Get it wrong and nothing above it matters Most teams never audit this split. They go straight to the copy and wonder why the reply rate doesn't move. The teams that consistently get replies aren't the ones with the best copywriters. They're the ones who got the offer right first and built everything else around it. Volume doesn't burn your TAM. Weak offers do. Apply for our no-cost pilot: https://bit.ly/C17Pilot Repost to help others close more sales. Follow Enzo Carasso 🧲 for outbound systems and revenue execution. | 60 comments on LinkedIn

The 7-Email Sequence that books +51% more meetings (3 Initial Emails + 4 Strategic Bumps) “Finally… this feels intentional.” That’s what a VP of Sales said after we rebuilt their outbound sequence. Before that, it looked like this: ● Random first emails ● “Just bumping this” follow-ups ● No logic behind the touches Here’s what actually works: 1) Trigger-Based Email ↳ This is the first email you send when you can anchor your outreach to a real trigger, hiring, funding, a tech change, a product launch, expansion, etc. ↳ The structure is simple: Trigger → Pain → Proof → Soft CTA. ↳ It works because it gives the buyer a reason to care now, not later. 2) Competitor Email ↳ This is when you know they’re using a competitor. ↳ You acknowledge what the tool is good at, then highlight the gap that usually forces teams to look elsewhere, and invite a conversation around that. ↳ It reframes without sounding like an attack. 3) Bold Observation Email ↳ This one is short and direct. ↳ You make a sharp observation about what they’re doing, point out the common pattern, and end with a curiosity-based close. ↳ No pitch. Just a call to conversation that gets them to self-diagnose. Then the bumps: 4) External Resource Bump ↳ Instead of “checking in,” you share a relevant third-party resource tied to their context. ↳ This borrows authority and makes your follow-up feel useful, not needy. 5) Data & Stats Bump ↳ Lead with a number that makes the problem measurable. ↳ Stats make pain tangible, and tangible pain gets replies. 6) Thoughtful Context Bump ↳ This is where you re-anchor urgency without pressure. ↳ You remind them why it’s relevant based on their situation, without sounding like you’re chasing. 7) Clarification Email ↳If they didn’t reply because they were confused, this fixes it. ↳You restate what you do in one line, clarify why it matters to them, and ask a clean alignment question. That’s the system. No begging. No random nudges. No volume increase required. If your cold email strategy sounds like begging, DM me “REPLY” and we’ll fix the sequence properly (let’s see if we’re a fit, and let’s work on your next campaign together) PS: Grab my FREE +10 cold email templates like +6847 sellers did already herehttps://lnkd.in/e4QY3TPW | 30 comments on LinkedIn

I send 50,000 emails per month and if i had to rebuild my outbound stack from scratch today, here's what I'd use. 1. Email infra --> I'll go with Zapmail --> Clean deliverability and decent pricing. --> I've never had a reason to second guess it. 2. List building --> Depends on the use case, but AI Ark, Apollo.io, Sales Nav cover most of it. --> If I needed niche data I'd scrape directories 3. Enrichment and Qualification --> Clay. No doubt on this one. --> Clay is whatever you mold it to. --> If you know how to build enrichment prompts in there, that's the skill. 4. Copywriting --> Currently using Claude. Just Claude. --> I've tried Gemini. It doesn't hit the same. 5. Sequencing --> Smartlead for email, GetSales.io for LinkedIn. --> I've used lemlist, Instantly.ai, Heyreach, etc. Are there better options? Probably. Do I care? Not really. I genuinely couldn't care less between this one or that one. They all do the same thing. Pick one. 6. Call recording --> Fathom.ai. Free, simple, does the job. --> I switched from Fireflies because it was free. What I wouldn't bother with? Claude Code. Bit of hype, not much ROI for what I actually do. P.S. If you wanna see how we write emails like no one else, and generate tons of meetings for our clients, book in a call here: https://lnkd.in/euDUSQyx | 12 comments on LinkedIn