AI meeting summary template
In short: Use AI for the first pass, but require every decision and action item to point back to the transcript and pass a human review.
A useful AI meeting summary is not a shorter transcript. It is a decision record: enough context to understand the outcome, clear actions with owners, and open questions that have not been disguised as decisions. Copy the template below into your note or prompt.
Copyable AI meeting summary template
Paste this structure above the transcript or include it in your AI instruction.
Meeting title:
Date and time:
Participants:
Purpose / context:
Short summary (3–5 sentences):
Decisions
- Decision:
- Rationale:
- Transcript reference / timestamp:
Action items
- Action:
- Owner:
- Deadline:
- Dependency:
Open questions
- Question:
- Decision owner:
- Information needed:
Risks or disagreements:
Follow-ups / next meeting:
Human verification
- [ ] Names and roles checked
- [ ] Decisions checked against the transcript
- [ ] Owners and deadlines confirmed
- [ ] Sensitive details reviewed before sharingIf a field was not discussed, write “Not agreed” instead of asking AI to fill the gap.
How to use the template
Provide the meeting transcript and ask AI to fill only what is supported by that source. Tell it to preserve uncertainty, distinguish proposals from decisions, and leave absent owners or dates as “Not agreed.”
Read the output once against the transcript. The review should focus on consequences rather than grammar: what was decided, who committed, and what will happen next.
What AI can extract well
AI is useful for grouping topics, finding candidate decisions, listing explicit action statements, and producing a short narrative summary. It is less reliable when speakers contradict themselves, decisions are implied, or an owner is suggested but never agrees.
- Repeated topics and agenda sections
- Explicit decisions and assignments
- Dates stated in the conversation
- Questions that remained unanswered
What a person must confirm
Confirm whether the final statement really superseded earlier discussion, whether the named person accepted an action, and whether sensitive details should be shared. These are social and organizational judgments, not transcription tasks.
Where MemoFlow fits
MemoFlow can record and transcribe the meeting, then generate a meeting write-up inside the same note. Keeping the summary next to the transcript makes the human verification pass faster and leaves the reasoning traceable.
Key points
- Summarize outcomes, not every exchange.
- Require source support for decisions and actions.
- Leave missing owners and dates visibly unfilled.
- Complete the human verification checklist before sharing.
Frequently asked questions
What should an AI meeting summary include?
Include context, a short summary, confirmed decisions, action items with owners and deadlines, open questions, risks, and follow-ups.
Should AI infer missing action-item owners?
No. If nobody explicitly accepted the action, mark the owner as not agreed and resolve it with the team.