Execution · step 5 of the Growth Engineering Loop
Execution: the work, done by agents.
The work done across every channel, on whichever models are best; deliberately the least special part; measured by the level.
Powered by: Agents. Execution feeds neither axis of the map. The Level grades it.
Agents do the work every working day, across your site, AI answers, social, outreach and paid. The models are replaceable parts; what surrounds them is not.
Where Execution sits on the map
Execution feeds neither axis of the map. The Level grades it.
Each step is graded on one scale: Not there · In your head · Written down · Wired in. The weakest step is the grade.
See the map and the five levels
What they do
Write, cut, publish, pitch and measure, every working day, across your site and AI answers, social, outreach and paid.
Why this is the commodity
Every competitor has agents now, and the models behind them change every few months. Execution is not where the advantage is, which is why we treat models as replaceable parts and put the effort into what surrounds them.
What keeps them from producing sameness
They read your Growth Graph before they write, they pass the distinctiveness checks before anything ships, and they never send, post or spend without your say.
You can watch them work
The day's work appears in Slack: what was made, what was refused and why, what needs you. Nothing happens in a system you cannot see into.
Where this step is put to work
- Get recommended by AI: Writes the evidence and places it on your site and on the third-party sources answers are built from.
- Outbound pipeline: Measures the prospect's own position first, then drafts one message in your voice around it.
- Social audience: Runs recognisable, repeatable formats in your voice.
Questions
Which models do you use?
Whichever performs best for the task, and we change them when better ones arrive. That is a maintenance detail, not a strategy.
Can our team use them directly?
Yes, and the programme exists to teach exactly that.