Choose one project and carry it through the month. Adapt the schedule to access and complexity. This is a learning plan, not a certification.
| Days | Work | Artifact | Completion check |
|---|---|---|---|
| 1–3 | Define user, task, goal, and existing evidence | Problem brief | Evidence and assumptions separated |
| 4–5 | Apply AI to one bounded workflow | Input/output example | Claims checked against sources |
| 6–7 | Review learning and baseline effort | Decision brief | Another person can follow the reasoning |
| 8–10 | Compare solutions and choose a test question | Options table | Simpler alternative considered |
| 11–14 | Build and inspect one journey | Prototype | Mocked behavior documented |
| 15–18 | Observe users or run clearly labeled practice evaluation | Observation/evaluation log | Limits and failures visible |
| 19–21 | Define outcome, quality, and cost measures | Metric contract | Denominator and window explicit |
| 22–25 | Make one evidence-based revision | Revised artifact | Affected behavior rechecked |
| 26–28 | Explain the work and request critique | Case study and short walkthrough | Contribution and uncertainty clear |
| 29–30 | Choose the next skill and repeatable workflow | Next-month plan | One specific next action |
Weekly reflection
What did I actually learn? [ ]
What did AI get wrong or leave out? [ ]
What did I change because of evidence? [ ]
Which part required another person's expertise? [ ]
What will I do differently next week? [ ]
End-of-month record
Completed artifacts: [links]
Real versus fictional/practice material: [ ]
Observed result and limitations: [ ]
Most important unresolved question: [ ]
Next skill to practice: [ ]
Workflow to repeat: [ ]
If you cannot recruit users this month, document that limitation and make recruitment a next step. Do not substitute AI-generated respondents for evidence of actual customer behavior.