Anyone can generate the screen. Someone has to own the experience.
Duplicate this page, then open Start here: the first week. Twenty minutes for you today. Nothing else needs to be read first.
Start. Little process, AI use is individual, nothing to fix yet. Reading order: 00, 02, 01, 04, 03, 07.
Fix. A process exists and AI arrived through tools nobody chose together. Reading order: 00, 02, 03, 01, 04, 07.
Part 00 tells you which one you are on. Read parts 05 and 06 when the 90 day plan calls for them.
| Part | The question | You fill in |
|---|---|---|
| 00 Where you are | Is there a process? Where is AI already, and who owns it? | The self check, then Inventory |
| 01 Production map | What are our steps, what does AI do at each, where must a human decide? | The map, the gates, the step cards |
| 02 Who owns what | Which hats exist at our size, who decides, where does design sit? | Hats |
| 03 Rules | What is allowed, with what data, gated by whom? | The one page policy, Inventory, the exceptions log |
| 04 Experience quality | Who guards the user when anyone can generate output? | High severity list, Defect log, the patterns file in your repo |
| 05 Habits and incentives | How does behavior change, and what do we reward? | The incentive check, the rituals calendar |
| 06 Measuring | Did it work? | Scorecard, Edit rate tally |
| 07 First 90 days | What happens week by week on our path? | 90 day plan |
The design lens: every part ends with an experience quality check. What can go wrong for the user here, and who notices.