Use AI to analyze routine execution logs…

When This Fires

• When a routine fails repeatedly and you can't identify the structural cause
• When you have 4+ weeks of routine execution data (daily logs, habit tracker, journal entries)
• When "willpower" explanations for routine failure feel insufficient
• Complements Periodically perform automated steps manually — maintain… (manual calibration) with the AI-assisted diagnostic for routine systems

Common Failure Mode

Insufficient data: feeding AI a single week of logs and expecting structural insights. Structural patterns require multiple cycles (4+ weeks minimum) to distinguish from noise. Also: logging only completion/non-completion without context variables (sleep quality, wake time, prior activities) that the AI needs to find correlations.

The Protocol

(1) Log routine execution for at least 4 weeks, capturing both completion status and context variables: wake time, sleep quality, prior activities, mood, energy level, any deviations from normal. (2) Feed the complete log to an AI with the prompt: "Analyze these routine execution logs across all cycles. Identify which context variables and deviations correlate most strongly with routine failure (missed elements). Look for cascade patterns where one deviation predicts multiple failures." (3) Review the AI's findings for structural fragility points: the variables that most predict failure. (4) Design structural interventions targeting the top 1-2 fragility points rather than generic "try harder" responses. (5) Re-log for 4 weeks after implementing structural fixes to verify they addressed the actual fragility.