Six layers, in the order they have to run.
Everyone now executes marketing with the same AI. Growth Engineering is the layer above that execution — the part that decides what the tools should produce, refuses what is not yours, and proves whether it worked. Each layer exists because a specific thing goes wrong without it.
- Growth Graph — It knows your business. One living model of your company that every agent reads before it writes, decides or spends.
- Direction — You make the big calls. The machine handles its own routine mistakes and asks you only where judgement is genuinely required.
- Distinctiveness — You do not sound like everyone else. Every output is measured against the default version a model would write, and against what your category already says.
- Experiments — Tested before you spend. Simulate to remove the weak ideas, prove the rest in the market, keep what wins.
- Agents — They do the work. Execution across your channels, on whichever models are best. Deliberately the least special part.
- Evaluation — Proof on every piece, and every decision. Checks before anything ships, and a verdict afterwards against what it was supposed to do.
Run apart, they are six tools. Run in order, they compound.
The graph makes the agents specific. Direction keeps a person on the calls that need one. Distinctiveness stops the output converging on the average. Experiments turn opinions into evidence, and Evaluation writes down what actually happened — which goes back into the graph. That loop is the reason month twelve does not start like month one.