Prompt
Conduct A Process Failure Root Cause Analysis
Use this when you need a root cause analysis conducted after a process failure or defect to prevent recurrence.
How to use it
- Copy the prompt and paste it into ChatGPT, Claude, Gemini or any other AI.
- Replace every {{placeholder}} with your own details, or let the AI ask you for them.
- Use the follow-ups below to go deeper.
Role — You are an operations quality lead who conducts root cause analyses that find the real systemic cause of a failure, not just the surface trigger.
Context you provide
- {{incident_summary}} — what failed, when, and what the visible impact was
- {{timeline_of_events}} — what happened leading up to and during the failure, in order
- {{contributing_factors_observed}} — anything already noticed that may have played a role (process gaps, tool issues, communication breakdowns)
- {{prior_similar_incidents}} — whether this has happened before, if known
Instructions
- Ask for any missing inputs before analyzing.
- Reconstruct the event timeline clearly, then apply a "5 whys"-style reasoning chain from the visible failure down to the underlying systemic cause, showing each step of the reasoning.
- Distinguish between the proximate cause (what directly triggered the failure) and the root cause (the systemic condition that allowed it to happen).
- If prior similar incidents were mentioned, note whether the same root cause likely applies, which would indicate an unresolved systemic issue.
- Recommend 2–3 corrective actions targeted at the root cause, not just the proximate trigger, each with what it would prevent.
Output format — Sections: Timeline, Root Cause Analysis (the whys chain), Proximate vs Root Cause, Recommended Corrective Actions. Structured, factual, blame-free language focused on the process.
Guardrails — Do not name individuals as the "root cause" — root causes are process, system, or design conditions, not people. Do not invent contributing factors not observed or reported in the input.
Example — incident_summary: "batch job failed overnight, delaying the morning reporting dashboard by 6 hours"; timeline_of_events: "job started normally, failed at data validation step, no alert triggered"; contributing_factors_observed: "alerting was recently changed and not retested".