Clean up
After the meeting
Bring the raw notes, a transcript, or an AI meeting summary; it doesn't need cleaning up first.
Same conversation as your plan and the discussion guide. Brand-new conversation? Paste the discussion guide Part 4 produced first, so the blameless language filter and your context carry over.
First paste — sort, don't write
Before anything final gets written, the notes are sorted into three lists: what we actually confirmed (stated as fact, not inferred), what we're still assuming (plausible, discussed, not confirmed), and what's still open (questions the meeting never answered). The blameless language filter applies to all three. Then it stops — you confirm or correct the sorting before anything else happens.
Here are my raw notes from the post-mortem meeting we just ran using the discussion guide from Part 4: [paste notes, transcript, or meeting summary here]
Before writing anything final, sort this into three lists:
1. **What we actually confirmed** — stated as fact in the notes, not inferred.
2. **What we're still assuming** — plausible, discussed, but not actually confirmed in the room.
3. **What's still open** — anything the discussion guide asked about that the notes don't answer at all (e.g. the meeting never got to Recovery). Don't guess at these — list them as open.
Apply the blameless language filter from Part 4 to all three lists. Stop after this and show me the three lists — I'll confirm or correct before you write the final summary. Second paste — the final version
Once the lists are right: a 150–200-word summary built only from what was confirmed, and an action-item table — task, owner, due date, and how you'll know it's done. Vague commitments get flagged and rewritten into specific, checkable items, and anything still open is carried forward so it doesn't get lost. Formatted to paste straight into a doc or wiki page.
These lists are correct [note any corrections here]. Now write the final version:
1. **A 150-200 word summary** — what happened, what we learned, using only the confirmed list, not the assumptions. Same blameless, system-focused language as before.
2. **An action item list**: task, owner, due date, and how we'll know it's done. If a commitment is vague ("we'll be more careful," "improve communication"), don't write it down as-is — flag it and propose a specific rewrite instead, the same way the Future Actions section in Part 4 does.
3. **Open items** — carry over anything still unresolved from 5a so it doesn't get lost.
Format as something I can paste straight into a doc or wiki page: headers, the summary paragraph, then the action item table, then open items.