Share completed reflective writing with AI and ask…

When This Fires

• After completing any significant reflective writing (journal entry, annual review, decision analysis)
• When reflection feels circular — reaching the same conclusions without new insight
• When you suspect your analysis of a situation has blind spots but can't identify them
• Complements Rate each critical assumption… (assumption testing for strategies) with the personal-reflection application

Common Failure Mode

AI-led reflection: opening ChatGPT and asking "Help me reflect on my week" produces AI-generated reflection that surfaces the AI's patterns, not yours. The human must write the reflection independently first — the writing IS the thinking. AI's role is analytical review of human-generated reflection, not generation of the reflection itself.

The Protocol

(1) Complete your reflective writing fully. Do not self-censor or anticipate AI feedback. Write what you actually think and feel. (2) Share the completed text with AI. (3) Ask specifically: "What assumptions am I making in this reflection that I haven't examined? What premises am I treating as facts?" (4) Review the AI's response critically — not every identified "assumption" is actually unexamined, and some may be deliberately held. But typically 1-2 of the surfaced assumptions will be genuinely unexamined and worth questioning. (5) For each genuinely unexamined assumption: "Is this actually true? What evidence supports it? What would change in my analysis if this assumption were wrong?" This is the productive inquiry that self-reflection alone couldn't reach.