See It Work — A maturity score the board will actually believe | Book 10 · Scaling AI Agents

Every maturity assessment you've run is a self-graded questionnaire, and the board knows it. Watch your maturity get read a different way — from what your agents have actually done.

A survey says what you hope. A receipt says what happened. The credible maturity number is the one backed by real, attested activity — impossible to inflate.

What this means for you

Maturity you can prove beats maturity you assert. When the score is computed from the attested actions your agents actually produced, it survives scrutiny — and it points you at exactly what's missing to reach the next stage. What this means for you: you can show the board a maturity score they'll actually believe, because it's computed from what your agents really did, not from a self-assessment.

The stage came from your real activity. Here's the evidence it was read from:

Maturity Evidence

distinct_action_classes5

functions_covered3

maturity_stage3 — Scaling

basisattested receipts, not survey

Recomputable by anyone with the receipts — which is what makes it a number leadership can stand behind.

For the technical reader — the command, and how to verify it yourself

# one line · you do not need to run this
see walkthrough

jq '[.[].action_class] | unique | length' memory/*_summary.json
# -> your distinct action-class count

Full step-by-step is in Appendix RX: Hands-On Demonstrations in the book.