Meeting Action Record
Turn meeting notes into a factual JSON record of decisions, questions, actions, dependencies, and commitments.
Customise & copy
Add the context you have. The prompt updates as you work.
Paste the raw transcript, notes, or call summary.
List attendee names and organizations if relevant.
State what the meeting was intended to accomplish.
Provide confirmed dates or milestone deadlines known before the meeting.
Map attendees or stakeholders to their roles and responsibilities.
Live prompt
5 fieldsConvert the meeting material below into a concise, factual JSON action record. Meeting transcript or notes: {{meeting_text}} Attendees: {{attendees}} Meeting agenda or objective: {{meeting_objective}} Known project deadlines: {{project_deadlines}} Stakeholder roles: {{stakeholder_roles}} Return only valid JSON with exactly these top-level keys: { "confirmed_decisions": ["..."], "unresolved_questions": ["..."], "action_items": [ {"owner": "...", "deadline": "...", "task": "..."} ], "dependencies": ["..."], "commitments": ["..."] } Rules: - Include only decisions explicitly confirmed in the transcript or notes; do not treat suggestions, possibilities, or discussion as decisions. - Capture every clearly stated action, with the named owner and deadline when available. - Use "Unassigned" for a missing owner and "Not specified" for a missing deadline; do not invent facts. - Include only dependencies and commitments explicitly stated or directly supported by the material. - Preserve dates and names as provided. - Use empty arrays when a section has no supported entries. - Do not add commentary, explanations, markdown, or extra keys.
Proof before you use it
A real example
Tested on gpt-5.6-luna on 2026-09-09
{
"attendees": "Jordan Lee — Product Manager; Priya Shah — Engineering Lead; Marco Ruiz — Customer Success Manager",
"meeting_text": "Jordan: We confirmed that the new onboarding flow will launch to 10% of users on June 14. Priya will merge the accessibility fixes by June 7. Marco will prepare the customer-support FAQ before the rollout. The analytics event naming is still unresolved because Data has not approved the final taxonomy. Jordan will ask Data for approval. The launch depends on the security review being completed first. Everyone agreed to review the rollout dashboard each morning during the first week.",
"meeting_objective": "Finalize the onboarding-flow rollout plan and identify launch blockers.",
"project_deadlines": "Security review: June 6; accessibility fixes: June 7; 10% rollout: June 14.",
"stakeholder_roles": "Jordan Lee — Product Manager; Priya Shah — Engineering Lead; Marco Ruiz — Customer Success Manager; Data team — analytics taxonomy approval."
}A small ritual that works
How to use it
- 01Paste the meeting transcript or notes.
- 02Add attendees and the meeting objective.
- 03Optionally provide project deadlines and stakeholder roles.
- 04Review the returned JSON before sharing or importing it.
Why it works
The prompt defines a fixed JSON schema with specific keys for decisions and actions, ensuring consistent output structure. It includes explicit rules that restrict output to confirmed information, reducing ambiguity in the generated record.
Where it fails
INSUFFICIENT_EVIDENCE
Customisation tips
Populate optional fields like project_deadlines to ground deadlines in known constraints. Ensure meeting_text is complete to support accurate extraction of action items.
Watch-outs
- Required variables must be provided to avoid missing context.
- Output must remain raw JSON without markdown formatting to satisfy parsing requirements.
Works in
Was this useful?
Submit your variation
If your version helps, it may be published with your name and a link back to you.
Related prompts
Want to make AI useful across your team? Explore Atul's training.