Meeting Notes Prompt — a prompt built for one specific pain: nobody wants to write the minutes after a meeting. Copy the prompt below into any LLM chat, fill in the 5 variables, paste your raw material (a transcribed recording, chat logs, or your own rough notes), and the AI will hand back a complete, archivable set of meeting notes — with action items, owners, and open questions.
When to use it
- After standups, project reviews, or requirement syncs where you only have scattered notes.
- You have a transcript of the recording but no time to clean it up sentence by sentence.
- You need two outputs at once: the official notes for everyone, and a to-do list for yourself.
The core value: it turns "writing up the notes" from an hour of manual work into one paste and two quick checks. It won't make judgments for you — anything not in the raw material stays out; rule 6 in the prompt enforces that.
Copy this prompt
# Role
You are our team's meeting-notes specialist, skilled at turning messy raw records into clean, archivable meeting notes.
# Meeting info
Meeting topic: {{Meeting topic}}
Meeting date: {{Meeting date}}
Attendees: {{Attendees}}
Output format: {{Output format}}
# Raw records
---
{{Raw records}}
---
# Rules
1. Start with a one-sentence summary of the meeting's goal and final conclusion;
2. Organize discussion points by agenda item, keeping key disagreements and the consensus reached — don't just record conclusions while dropping the debate;
3. Extract every action item, each with: task, owner, due date; if the source didn't name an owner or date, mark it "TBD" — never invent one;
4. List unresolved issues separately as "Open questions", noting where each is stuck and who needs to decide next time;
5. Remove filler, small talk, and repeated remarks; don't merge different people's views, and don't invent anything not present in the raw records;
6. If key information is missing (e.g., an agenda item only has half a sentence), end with an "Information needed" list asking me about each gap.
# Output structure
Output in Markdown with four sections: Meeting overview, Agenda notes, Action items, Open questions.What each of the 5 variables takes
| Variable | What to fill in | Example |
|---|---|---|
| {{Meeting topic}} | Meeting name, one line | Q3 growth review |
| {{Meeting date}} | Date of the meeting | 2026-09-26 |
| {{Attendees}} | People present, comma-separated | Anna, Ben, Chris |
| {{Output format}} | How you want it presented | Markdown notes / Table of action items / Email body |
| {{Raw records}} | Transcript, chat log, or rough notes — paste directly | ... |
5 tips to get more out of it
- Feed long meetings in chunks: for meetings over an hour, generate notes per agenda item first, then ask the AI to merge them — more reliable than pasting 20,000 words at once.
- Notes first, to-dos second: after you get the notes, follow up with "turn the action items into a tracking table with task, owner, due date, and status" — paste it straight into your project tool.
- Change the format without changing the prompt: only the {{Output format}} variable needs to change — the same prompt produces both the "official version for everyone" and the "minimal version for yourself".
- Redact before pasting for sensitive meetings: replace names or figures tied to compensation, layoffs, or unreleased data with placeholders first; or use your company's internal model.
- Always review the action items yourself: the AI occasionally mistakes "someone mentioned it" for "it was decided" — two minutes checking owners and dates before sharing is the one step you can't skip.
If you also want the AI to play a fixed role in your workflow, this site's turn AI into your mock interviewer: how to write role-setting prompts uses the same logic: define the role first, then the rules, then constrain the output.
Once you have this meeting notes prompt in hand, all you need to do in the meeting is listen and jot things down — one paste handles the rest.
