Hey, it's Jose.

This issue is different from the usual.

Attio published something called GTM Atlas last month. Fourteen operators, from Lovable, Vercel, Framer, Notion, Linear, Stripe, Granola and Anthropic, writing about how they actually run go-to-market right now. Ungated. No form, no email wall, no "book a demo”.

I read the whole thing in one sitting, and then went back through it a second time with a notebook.

Here's the part I want to be honest about, because it changes how you should read it. The Atlas is written for greenfield teams. Attio says so directly: the first GTM hire, the founder building revenue infrastructure, the RevOps hire starting with no legacy constraints.

That is probably not you.

You have an existing operation. A CRM someone configured three years ago and then left. Outbound plans that started life as a Google search. Five systems that report five different numbers for the same metric, and a standing meeting where people argue about which one is right.

I believe that makes the Atlas more useful to you, not less. Reading advice written for people without your constraints is the fastest way to find out which of your constraints are actually real, and which ones you adopted and never questioned.

Here are the five things I'd act on.

Kyle Norton runs GTM at Owner.com and calls this the hill he'll die on. Outbound has one point of failure, and it's the data. Tier your list, if you’re honest, you'll find roughly a 25% is A, 50% is B, and a 25% is C (Tier C is who you will never close). Reps hit the activity number by working Tier C. Everybody feels busy. Nothing closes.

This is the Intelligence Layer problem in its plainest form, and it's why the framework opens there: you don't have a data problem, you have an access problem.

Do this week: tier last quarter's outbound list against closed-won, then calculate what share of rep hours went into Tier C. That percentage is your business case for everything below.

Same entry, sharper point. Most teams think the unlock is giving every rep their own assistant to build lead lists. Kyle argues the opposite: one specialist owns the build, gets it to production quality, and ships one version to everyone. His reasoning is that expert-versus-novice doesn’t get you 50% more, an expert is orders of magnitude better.

And the money is already on your P&L. A GTM engineer or an expert consultant costs roughly two BDRs. Take two BDR slots and invest in one.

This is the GTM Architect pillar with a price tag attached. RevOps stops being the queue that fixes what other people broke and becomes the function that builds what scales.

Do this week: find the rep who's already quietly building their own workflows. Either promote that work to the standard or shut it down. 12 separate unique workflow versions is not enablement, it's technical debt with a quota.

Maja Voje's entry is the most usable artifact in the collection. Her point is that most teams use AI like a chatbot. The model never remembers your ICP, your positioning, or what you decided last quarter. She lays out a persistent structure instead:

A team still prompting is running a chatbot. A team doing this is running a “GTM Brain”. That distinction is the whole intelligence layer argument in one line, and it costs a weekend rather than a budget cycle.

Do this week: write the standing context file. One page. ICP, positioning, current priorities. Nothing else works without it.

Rati Zvirawa at Intercom tells a story about a customer convinced that an AI agent had broken their MQL numbers… This is until they discovered the MQL definition was not consistent across funnels. The agent didn't break anything. It just surfaced what was already broken.

Emily Kramer goes further and argues the MQL/SQL framework itself creates territorialism. “The buying journey is not linear… It’s just false that someone’s handled by marketing and then handed off to sales,” she says.

We talked about this within the GTM Reloaded Agentic Workforce pillar. Agents don't repair a broken operating model, they publish it at scale.

Do this week: ask 3 people in 3 functions to write down their definition of a qualified lead, separately, without conferring. Compare. You now have, for free, what an AI deployment would have found the hard way.

Josh Epstein at Coder shares the number I liked most in the entire Atlas: 70% of licenses deployed across 80% of customers within 90 days. Customers who went through their formal technical validation hit it every time. Customers who skipped it hit it about 40% of the time. He also notes that churn is visible six months out, and it shows up as silence rather than complaints.

Silence is the hardest signal of all to capture, which is the argument for the Intelligence Layer stated from the retention side rather than the pipeline side.

Do this week: measure day-90 deployment for every customer you closed last year, and plot it against who renewed. That chart will tell you more than your health scores do.

Both Nicolas Sharp, Attio's own Founder, and Travis Bryant, who runs mid-market GTM at Anthropic, say the same thing there's no universal playbook.

The playbook was always bespoke, and what AI changed is what the human inside it does all day.

That's the whole reason I started GTM Reloaded. Not to hand you the playbook, but to show you what good has looked like so you can build the version that best fits your business.

One ask: which of these five would break something if you tried it at your company?
Hit reply. That's usually where the real issue is hiding, and it's what I'd rather write about next.

Talk soon,

Jose Celorio Founder, GTM Reloaded
Former Strategist at Google, Mastercard & Deloitte Consulting

GTM Reloaded is free. If you found this valuable, forward it to a colleague who's building a GTM operation that punches above its weight.

Not subscribed yet? Sign up here →