3 revenue motions your AI is only half wired into
Pipeline, demand, and deals. Where AI actually amplifies each, where most lean teams have it half-connected, and the 4-week play to embed 1 motion end to end.
Save this. 3 AI moves from this week’s Claude Signal, each with my operator take and a move a lean European team can run today, plus the play behind it.
Send it to 1 founder or GTM operator scaling toward 10M in Europe who reads everything and ships none of it.
The 90-second version
The model race tied. OpenAI shipped a Cowork clone, and Charlie Hills’ 5-task bake-off came back 4 ties. Buying a better AI no longer moves your number.
So the question flipped. Not which AI, but how deeply it is wired into the revenue motions you already run: pipeline, content, deals.
The waste is AI that never touches pipeline. Ruben Hassid: the real cost is paying for AI your team does not touch, which for a revenue team is pipeline left on the table.
You underproduce demand. If every asset runs on your most expensive model, cost caps how much demand you can create. Plan expensive, produce cheap, and reinvest the saving in volume.
The European read: a lean team spread across markets has no spare seats to waste and no second AI budget, so the edge is wiring 1 motion deep, then the next, and your winning-deal context is the part a competitor’s model has never seen.
Read time: 7 min
This week’s number 1 move
The AI stopped being the edge. Where it runs in your revenue motions is.
Two things happened this week and they point the same way. OpenAI shipped ChatGPT Work, a near-copy of Claude’s Cowork, and Charlie Hills ran both against Claude on 5 real build tasks and got 4 ties. When the top models draw even, buying a better AI stops moving your number. What moves it is how deeply AI is wired into the revenue motions you already run: pipeline generation, your content engine, your deal follow-ups. Ruben Hassid put a number on the waste. Past 150 seats, usage tiers can run 17 to 37 thousand dollars a month, but the real cost, in his words, is paying for AI your team does not touch. For most GTM teams that is exactly what is happening: reps with AI seats still building pipeline by hand.
My take: this newsletter runs on embedded AI, not on a better model, and that is the whole point. The edge is the wiring, not the tool I use. A lean team cannot out-buy a bigger one on models, but it can out-embed one, and that is the cheaper edge this year.
Do first: take your top pipeline-generating motion, list its 5 steps, and mark each one AI-run or still by hand. The by-hand steps are your embed list.
The rule: the AI is no longer the edge, where it runs in your revenue motions is. Wire 1 motion, then the next.
Sources: Ruben Hassid, “ChatGPT shamelessly copied Claude” (Jul 15) · Charlie Hills, “Claude 5 vs ChatGPT 5.6” (Jul 12)
Hi, it is Koen Stam and welcome to GTMcraft: The Future GTM Operator. This newsletter is built from 100,000+ GTM signals collected from 100+ operators and founders, combined with 13+ years of my own lessons and failures from the trenches. I write at the intersection of go-to-market practice and AI-powered systems for founders scaling 0 to 10M+ ARR.
100,000+ GTM relevant signals from LinkedIn, Newsletters and Podcasts indexed. Translated into 100+ GTM plays for you to implement today.
13+ years over 3 GTM operator jobs across 3 GTM motions (SMB, MM, ENT). Scaling from 2-10M+ ARR multiple times. Same recipe. Different motions.
Now all part of GTMcraft OS.
Problem 1: Your reps have AI seats, but your pipeline still gets built by hand
You bought AI seats for the revenue team. A couple of champions use them, and the rest build pipeline the old way: manual account research, hand-written first touches, call prep from scratch. Ruben Hassid ran ChatGPT’s new Work mode against Claude for a week and landed on the line that should reframe your GTM budget: the real opportunity cost is not Claude versus ChatGPT, it is paying for AI your team does not touch. On a revenue team that waste is not abstract, it is the pipeline those dark seats never generated. And Charlie Hills’ 5-task bake-off says the models have tied, so switching vendors adds nothing if the seat stays closed on the work that fills the funnel. The pipeline only shows up when a rep points AI at account research, list-building, and first-draft outreach instead of doing it by hand.
My take: this is the same call I make with any tool I pay for. A seat that never touches pipeline is lost pipeline, not just wasted spend, and that is far more expensive. I would rather have 3 reps who actually embed AI in their outbound than 10 seats sitting dark on the exact work that fills the funnel.
The European read: a lean multi-market team cannot carry dark seats. Every rep runs their own country, so an AI seat that never touches pipeline in that market is a hole nobody else backfills. Adoption inside the motion is the lever, not the vendor.
Do first: list your revenue reps and mark who used AI on a pipeline task this week, research, list-building, or outreach drafts. The dark seats are your pipeline leak.
Do this week: take the 1 outbound step reps still do by hand and make AI the default way it gets done, with 1 shared prompt. Ask your 2 lowest-AI reps the single blocker keeping them out, then fix it or reassign the seat.
Do this month: set a monthly check on AI use per revenue motion, and treat a seat that never touches pipeline as a cost to coach or cut, not a sunk one.
The rule: if a paid AI seat never touches pipeline, it is lost pipeline, not just wasted spend. Embed it in 1 outbound step or reassign it.
You know it worked when: more of your first-touch research and outreach drafts start in AI, and rep usage climbs toward the seats you pay for.
The play: 5 steps to turn AI seats you pay for into AI your team uses (2026 series). Count the dark seats, cut duplicates, activate the motion, and sustain adoption with a monthly check. Get full access below ↓
Sources: Ruben Hassid, “ChatGPT shamelessly copied Claude” (Jul 15) · Charlie Hills, “Claude 5 vs ChatGPT 5.6” (Jul 12)
Problem 2: You underproduce demand because every asset runs on your most expensive model
Demand creation is a volume game, and most of it lands on the 95% of your market that is not ready to buy yet, where you earn mindshare long before the deal. But if every campaign, sequence, and repurposed post runs on your most capable model, production cost caps how much demand you can afford to make, so you create less than the market needs. Charlie Hills shows the fix: plan the campaign and the positioning on the top tier, then produce the volume on a cheaper model, and review on the top tier. Same output, a fraction of the cost. Ruben Hassid shows the stakes, usage-based pricing spikes once your content volume crosses a threshold, so an unrouted engine gets more expensive exactly as you try to scale demand. Route the cost down and a lean team can afford to create demand across the whole market, not just the in-market 5%.
My take: I run this split on my own demand output, top model to set the angle, cheap model for the volume, and the saving goes back into producing more, not into a lower bill. On a lean European budget, that is the difference between building mindshare across the whole market and only touching the buyers already in a deal.
The European read: a lean team in fragmented markets has to create demand in several languages at once. Cheap production is what makes multi-market demand creation affordable, so you build mindshare country by country instead of rationing it to 1.
Do first: list your 5 highest-volume demand and content tasks this week and mark each as strategy, like positioning and campaign angle, or execution, like variants, repurposing, and enrichment.
Do this week: route the execution tasks to a cheaper model, keep the top tier for the strategy and the core message, and put the saving into producing more demand, not a lower invoice.
Do this month: write the plan-expensive, execute-cheap split into your content SOP, and set a demand-output target, pieces shipped per week, so the saving turns into reach, not just margin.
The rule: plan the campaign on the expensive model, produce the volume on the cheap one, and reinvest the saving into more demand.
You know it worked when: you ship more demand per euro, reaching more of the 95%, while reply and conversion rates hold.
The play: 5 steps to build demand with the 95% who are not ready to buy (2026 series). Create demand across the out-of-market majority, and produce it cheaply enough that a lean team can sustain the volume. Get full access below ↓
Sources: Charlie Hills, “Claude 5 vs ChatGPT 5.6” (Jul 12) · Ruben Hassid, “ChatGPT shamelessly copied Claude” (Jul 15)
Problem 3: Your AI drafts every deal touch, and they all read generic
Your reps let AI draft the discovery recap, the follow-up, the mutual action plan, the deck. The problem is every AI has read the same internet, so the drafts sound like every other vendor’s, and buyers feel it. Charlie Hills’ bake-off makes the point from the winning side: the models tied on almost everything, and the 1 task Claude took outright was the pitch deck, because it pulled real brand context instead of defaulting to corporate stock-photo energy. That is the whole game in a deal cycle. The differentiator is no longer the model, it is the context you feed it: your ICP, your last 10 closed-won calls, your objection handling, the language your champions actually use. Loaded with that, AI drafts a follow-up that sounds like your best rep on their best day. Blank, it drafts a template.
My take: the context is the part a competitor’s model has never seen, and it is the one thing I will not hand over blank. Polished is the floor now, everyone has it. A follow-up that carries our actual discovery, our value framing, our proof, is the edge, because no model has my team’s last 10 wins.
The European read: across markets, the thing a competitor’s agent cannot copy is your own winning-deal context, in the language your buyers actually use. Load it once and every rep in every market drafts from your best deals, not from the generic internet.
Do first: take your last AI-drafted follow-up and ask 1 question: could any competitor have sent this, or does it carry our discovery, our value framing, our proof.
Do this week: pull your last 5 closed-won deals, the call notes, the winning message, the objections you beat, into 1 context file your reps’ AI drafts against. Add a check to the prompt: does this reference the specific pain, metric, and next step from this deal, or is it generic.
Do this month: build a living deal-context file from every win and loss, so the AI your reps draft with gets sharper each quarter instead of staying generic.
The rule: the model tied, so context breaks the tie. If a competitor could have sent the same AI draft, load your winning-deal context before it goes out.
You know it worked when: reps stop rewriting AI drafts from scratch because the first draft already sounds like your best deal.
The play: 5 steps to make your AI deal drafts sound like your best rep (2026 series). Build a winning-deal context file, wire it into the draft, and check every touch against it before it goes out. Get full access below ↓
Sources: Charlie Hills, “Claude 5 vs ChatGPT 5.6” (Jul 12) · Ruben Hassid, “ChatGPT shamelessly copied Claude” (Jul 15)
Steal this move: embed AI into 1 revenue motion end to end in 4 weeks
Play: 5 steps to embed AI into one revenue motion end to end (2026 series)
Fixes: reps who have AI seats but still run the revenue motion by hand, so the AI you pay for never touches the number.
Best for: a founder or operator at 0 to 10M+ ARR paying for AI seats and running pipeline, content, or deals where AI is only half wired in.
Week 1, Define: pick your highest-value revenue motion, outbound, demand gen, or deal follow-up. Map its steps and mark each AI-run or by-hand. Pull who on the team actually uses AI in it today.
Week 2, Build: make AI the default on the 2 biggest by-hand steps with shared prompts. Load your winning-deal or brand context into 1 file the motion drafts against. Mark each step strategy or execution.
Week 3, Validate: route the execution steps to a cheaper model for a week. Compare output, cost, and the motion’s own metric, replies, meetings, or conversion, against the baseline.
Week 4, Operate: write the motion’s AI SOP, which model, which context, which human check. Set a monthly adoption-and-cost check. Then repeat the play on the next revenue motion.
Paste this into your AI:
Here is 1 revenue motion I run and its steps [paste]. Mark each step AI-run or by-hand and strategy or execution, tell me the 2 by-hand steps to embed first, which model each step should use, and what winning-deal or brand context to load so the output sounds like us. End with the 1 metric I should watch to prove it worked.
Full play, template and workbook inside GTMcraft. Reply to this email or DM me to get access.
Also on the radar
Ruben Hassid, “ChatGPT shamelessly copied Claude” (Jul 15): ChatGPT Work now clones Cowork, so half your reps may drift to one tool and half to another, splitting your GTM context and doubling spend. Name 1 primary AI surface for the revenue team before the split hardens.
Ruben Hassid, “ChatGPT shamelessly copied Claude” (Jul 15): usage-based AI pricing spikes past a threshold, 17 to 37 thousand dollars a month past 150 seats on one vendor while another stays flat per seat. Put next year’s rep count into a 1-line AI cost model per vendor before you scale the team.
Charlie Hills, “Claude 5 vs ChatGPT 5.6” (Jul 12): Claude won the deck on brand context, not raw capability, and the same holds for QBRs and renewals. Load the account’s usage and wins into AI before it drafts the next renewal or QBR.
Why this matters now
For 2 years the GTM question about AI was which model is best. This week that closed. OpenAI copied Cowork, the top models tied on real work, and buying a better one stopped moving the number. The edge is now embedding: which of your revenue motions actually run on AI, how cheaply you produce the volume, and how much of your winning-deal context the AI carries. This extends the routing thread from Write The Loop on July 17. Last week the routing lived inside 1 tool, this week it goes cross-motion, into pipeline, content, and deals.
For a lean European team scaling toward 10M: embed 1 motion, route the model, and load your context. The founders who win the second half will not have the best AI. They will have AI wired deepest into pipeline, content, and deals.
What is your number 1 takeaway this week? Reply or send me a DM.
Go deeper with GTMcraft OS
Scale on plays, not prompts.
You do not need another newsletter, podcast or LinkedIn post. You need the system under it.
Read the plays. Every play, live and current, inside GTMcraft OS.
Download for your archive. Take the PDF or workbook offline.
Get the markdown. Drop the raw play into your own stack or AI tool.
Build your own. Use the GTMcraft Notion databases (100,000+ signals) to spin up a play for your exact motion.
The full Plays database is over 125+ plays and growing every week.
Curious to see a sneak peek of GTMcraft OS, DM me or reply on this email.
See you inside GTMcraft OS, your GTM co-operator. Not another community.
Koen
Your (human) GTM Agent
