Most teams know they are somewhere between “we have Copilot” and “the agents do it”, without being able to say where. This puts a number on it, scored on two axes, with the next concrete step for each. Your result appears on the page. No email needed to see it.
Most AI maturity assessments give you one number, and one number hides the thing worth knowing. How you build and what you sell move independently. A company shipping an AI product on a team with no shared system is fast and fragile. A team with an excellent internal setup selling traditional software is efficient and exposed. Averaging those into one score tells you nothing, so we score them separately. The framework behind it is in the four levels of AI maturity.
Committed alongside the code: a CLAUDE.md or equivalent, agent definitions, saved prompts, custom skills.
Past decisions, domain rules, why things are built the way they are.
The question behind it: when the next model ships, does your advantage disappear or compound?
Answer all 8 to see your reading. Nothing is sent anywhere until you ask.