Eight-question diagnosis · Chapter 13

Do not count tools. Test whether the organization can judge.

These questions test judgment and action when nobody is watching—not stated knowledge. No personal information is required, and answers are not sent to analytics.

Select only items verified by observed behavior

Do not score self-reports. Ask three randomly selected core members separately, use the lowest score, and treat conflicting answers as a failure.

01Can core members explain the eight-layer standards in plain language?
02Can the last three important decisions be traced to explicit value judgments?
03Does everyone have a stage-appropriate ROI card whose leading indicators support an ROI above 1 by the deadline?
04Is each core member's decision authority level—and the condition for escalation—clear?
05Does daily reflection consistently produce new standards, revisions, or better solutions?
06Do people and AI act from the same context and apply the same standards?
07Does product development share one user consensus, one product path, and one uncertainty under test, while preserving a stable core and rollback plan?
08When experience conflicts with a standard, do members pause, seek evidence from both sides, and revise the judgment instead of appealing to seniority or habit?