Management & Strategy · Updated July 17, 2026
ChatGPT prompt for running a blameless postmortem
This is the prompt we run within a week of a failed project, a lost client, or an outage, while memory is fresh. Its core rule: no individual blame. If your account blames a person, the prompt reframes it toward the system that let it happen. You fix the cause, not the culprit.
The prompt to copy
Run a blameless postmortem on: {what happened}. Facts and timeline as I know them: {paste}.
Produce:
1. A clean timeline from my facts only — mark gaps [UNKNOWN] rather than filling them.
2. Contributing causes grouped by process / communication / tooling / assumptions. No individual blame — if my account blames a person, reframe it as the system that let it happen.
3. What worked and should be kept.
4. At most 3 concrete changes — each with an owner and a "how we'll know it stuck" check.
5. Questions to ask the team where my account is likely one-sided.
Replace the braces with what happened and your facts, even partial ones. The prompt sorts what you know from what is missing and from what is only interpretation.
When to use it
Within a week of a failed project, a lost client, an outage, or a bad launch, while memories are sharp. Early enough that the facts are fresh, distant enough that you are not settling it in the heat of the moment.
What you get back
- A clean timeline built from your facts alone, with gaps marked [UNKNOWN].
- Causes grouped by process, communication, tooling, and assumptions, never by person.
- What worked and is worth keeping, often forgotten in a postmortem.
- At most three changes, each with an owner and a “did it stick” check.
Real example: a lost client → a usable postmortem
Input: Brightside Studio, an 8-person agency, has just lost the client Fernhill Coffee. Maya, who runs it, pastes her facts, colored with frustration:
we lost Fernhill Coffee. Lea delivered the campaign late, again.
i think the client had said it needed to ship before their trade show, but i can't find where that was written.
Output, the part that matters:
Timeline: initial brief (date known) → trade-show deadline mentioned [UNKNOWN: through which channel, on what date?] → late delivery → client left.
Contributing causes (process): the critical trade-show deadline was recorded in no shared tool. The delay is not “Lea delivered late,” it is “a date the client was counting on lived only in a conversation, with no milestone and no alert.” Anyone in her seat could have missed it.
Notice what the AI did: where Maya’s account blamed Lea, the postmortem reframed it toward the system, a deadline never recorded, that the next project can fix. And it marked [UNKNOWN] instead of inventing which channel the deadline came through. That gap becomes the first question to ask the team.
This prompt is part of the library that ships inside the kit, with its usage sheet.
Frequently asked questions
- Does this prompt work with Claude or Gemini?
- Yes. The blameless reframing works identically in Claude, Gemini, or Le Chat. Paste the prompt and your facts and the behavior is the same.
- Why blameless instead of finding who was responsible?
- Because a postmortem that names a culprit teaches everyone to hide their mistakes next time. By reframing toward the system that let it happen, you fix the cause instead of punishing a symptom.
- What does the prompt do if my timeline has gaps?
- It marks them [UNKNOWN] instead of filling them in. A postmortem built on invented facts is worthless; the flagged gaps become the questions to ask the team.