The method in one page

1. Start with the business, not the tool.
2. Find one costly friction.
3. Map the capability around it.
4. Name current and appropriate maturity.
5. Score impact, plausibility, adoption, evidence speed, cost, and risk.
6. Stop immediately at zero impact or zero adoption viability.
7. Design the smallest useful experiment.
8. Agree on evidence and guardrails before starting.
9. Use real work and representative users.
10. Stop, continue, invest, or industrialize.
11. Manage decisions as a portfolio of outcomes.
12. Ask what should stop.

This is not a one-time AI exercise. It is an operating discipline for making business choices while the technology keeps changing.

The tools will change. The need for judgment will not.

---