GTM Atlas — Build the system before the message

Author: Roniesha Copeland, VP Sales, Strategic Accounts at Vercel
Source: atlas.attio.com/build-the-system-before-the-message
Date: May 6, 2026

Who is Roniesha Copeland

Over a decade building and scaling GTM teams for developer-led platforms. Most of that time at Stripe, then ran sales at v0, now VP Sales at Vercel.

Core Thesis

Outbound is a system with three parts:
- ICP — picking the right people to go after
- Narrative — a compelling message that gets them interested enough to engage
- Timing — sending it via a channel and at a moment where it resonates

AI supercharged the system: data layer is now truly actionable, timing is precise, and "narrative" actually means matching your message to a moment.

The instinct for an early team is to start with narrative. But you need to lay the data foundations for ICP and intent first, or you'll spend the next year retrofitting a system built to acquire the wrong customers.

The ICP Stack

Start with ICP. If you don't know yours, build a hypothesis around it. If PLG: you can see the pattern of people who use your product. The question becomes: what pain am I solving, for which type of organization, and who inside it is making or influencing that decision?

The ICP stack has four layers (bottom to top):

  1. Company + Persona — base. That's your fit layer.
  2. Intent — the layer AI has unlocked. Telemetry says a lead at a growth-stage startup has been posting about the issue you solve → higher intent score.
  3. Revenue — not a filter, a multiplier on effort. Fit + intent tell you who to pursue. Revenue tells you how much to invest.

The PLG Altitude Problem

With v0 (Vercel's generative UI app builder), they had a hypothesis: the best way to experience the product is the product. Custom-built an application for their ICP at each target company. Open rates were very high. Conversions were zero.

Why: altitude. In a startup, the buyer and the decision-maker are the same person. As you target larger companies, those roles get further apart. At enterprise scale, you're trying to reach a senior decision-maker who owns the budget and doesn't care about the product — they care that the product solves a business problem. Leading with product puts you at the wrong altitude.

PLG teams often get this wrong. Developer love and end-user awareness are real goals, but community and marketing will drive that better than outbound. Outbound should leverage what you learn from PLG to aim higher.

Enthusiasm Is Not a Buying Signal

Most early teams are excited that other people want to buy their product and don't pressure-test that excitement. That is not qualification.

You need to know:
1. They have a pain point you can solve
2. It's painful enough to solve now

If those two things aren't there, the deal either doesn't exist or doesn't close in a time horizon that makes sense. You'll just have pleasant conversations with people who love talking about your product, and nothing will close.

Treat Everything as an Experiment

You have a hypothesis, structure activities around it, decide in advance how you're going to evaluate the outcome, then decide whether to keep investing or move on.

When you work that way, things feel less trapdoor. You're less committed to a thing you've decided has to be absolutely right. Test personas, test messaging, test tactics.

The teams that do this well are always running something and willing to kill the thing that isn't working.