Map the decisions that matter.
Every program begins not with a model but with a map. List the consequential decisions the organization makes — the ones that move money, risk, or trust — and describe how each is made today: what triggers it, who owns it, what evidence stands behind it. Most teams discover the evidence is thin and the ownership is fuzzy. That discovery is the point.
You cannot govern what you have not named. The map is the contract for everything that follows.
Measure the before.
For each decision worth automating, capture a baseline before you touch it. How accurate is it today, how fast, how costly when wrong? Skip this and every later improvement is an assertion. Capture it and you have the one thing an auditor and a board both want: a measured delta.
- 01 — Applied BI. Make the decision visible before you automate it.
- 02 — Evals-first. Build the harness that proves it keeps working.
- 03 — Closed-loop value. Prove the value, do not just claim it.
- 04 — Governance by design. Compile the gates into the runtime.
- 05 — Sovereign by default. Start from the boundary the mandate requires.
Run it as an operating system.
The principles are not a checklist you complete once. They are how the program runs: gates that hold on every decision, a harness that fires on every change, a ledger that records every action, value measured in a closed loop. At that point you are no longer running a project — you are running AIIX Intelligence Runtime, and the organization has become measurably more intelligent.
The playbook ends where the operating system begins. Everything before it is preparation; everything after it compounds.
A playbook you run once is a project. A playbook compiled into the runtime is an operating system.