Skip to prompt
SecondBrain Add to Chrome
PromptsMenu

Productivity · Meetings

Meeting Notes to Action Items Prompt

  • Project decisions
  • Client follow-ups
  • Weekly operating reviews
  • Evidence-based records

Convert raw meeting notes into a structured, reviewable draft of decisions, action items, owners, dates, risks, and unanswered questions.

Free · No signup · Model-agnostic structure

Editorially reviewed by
SecondBrain
Updated
Setup time
4 min
expected_output.md

THIS PROMPT BUILDS

  • A concise meeting summary
  • Decisions with supporting context
  • Action items with owners and dates
  • Open questions and follow-up draft

Planning assistants

ChatGPTClaudeGeminiGrokCopilot
Jump to a section

Keep going only when you need help

The prompt is ready. Use these guides to improve the result.

What this is for

Use it when the output needs to hold up after the chat ends.

Meeting summaries often sound polished while quietly changing what people agreed. This prompt preserves uncertainty, separates decisions from discussion, and refuses to assign owners or dates that are not present in the notes.

Best for

  • Project and product teams
  • Client and stakeholder meetings
  • Weekly operating reviews
  • Research interviews and workshops

You will get

Decision record

Confirmed decisions are separated from proposals, with rationale and current status.

Action register

A structured table exposes the deliverable, owner, deadline, dependencies, and confidence.

Follow-up message

A concise attendee recap that can be verified before it is sent.

Worked example

Example: a product scope meeting

See how specific inputs change the shape and usefulness of the response.

Example inputs

Meeting purpose
Decide the Q3 onboarding scope
Team context
Product, design, and engineering leads
Date convention
YYYY-MM-DD; meeting date 2026-07-18
Raw meeting notes
Raw notes with speaker names and timestamps
representative_output.txt
  • Decision: ship guided setup and defer team invitations; rationale is engineering capacity; status confirmed.
  • Action: Maya to deliver the revised flow by 2026-07-22; dependency is final copy; confidence high.
  • Needs confirmation: no owner was recorded for analytics instrumentation.
Why this output is useful—and what to verify

The output distinguishes a confirmed decision from an unresolved assignment and preserves the dependency. A meeting attendee should still verify names, dates, and whether ‘defer’ was a final decision.

Practical guidance

Keep the notes and the summary connected

A useful action list makes missing information visible instead of smoothing it away. Keep the original notes available, record the meeting date, and make it clear when a person, deadline, or decision is only proposed.

The fast start is useful for turning a short note into a reviewable draft. Use the full brief when speaker roles, date conventions, or sensitive project context change how a statement should be interpreted.

Practical guidance

Review commitments before they become the record

AI can reorganize what was said, but it cannot establish consensus that is absent from the notes. A responsible attendee should compare each decision, owner, and deadline with the source before the recap goes out.

When the notes are incomplete, send the open questions as questions. Do not turn a plausible inference into an assignment just because it reads cleanly.

Step-by-step workflow

How to use this prompt well.

  1. 01

    Clean the input safely

    Remove sensitive details, label speakers only when roles matter, and include the meeting date for relative deadlines.

  2. 02

    Review the evidence flags

    Check every decision, owner, and deadline against the source notes before sharing the result.

  3. 03

    Confirm gaps with the team

    Use the open-question section to assign missing owners, resolve ambiguous dates, and close the loop.

Prompt anatomy

Why this structure works.

It protects the source of truth

Explicit non-invention rules reduce the chance that a polished summary silently becomes a false record.

It exposes missing commitments

Unassigned owners and unspecified dates become visible follow-ups instead of disappearing from the summary.

It separates decisions from discussion

The structure prevents a suggestion or debate from being rewritten as team consensus.

Before you use the result

Run a quick quality check.

A fluent answer is not automatically a correct or responsible one.

Ready-to-use checklist

  • Every decision is explicitly supported by the source notes.
  • Proposals, debate, and unresolved questions are not presented as consensus.
  • Missing owners and dates are labeled rather than guessed.
  • Relative dates are normalized only when enough context exists.
  • A responsible attendee has reviewed the record before distribution.

Better inputs, better judgment

Small choices that improve the result.

Make it stronger

  • Include the meeting date whenever notes use relative dates such as tomorrow or next Friday.
  • For long transcripts, summarize in chunks and then run the prompt on the combined evidence—not on summaries that omit nuance.
  • Keep a link to the original notes so reviewers can verify disputed decisions.
  • Do not paste confidential transcripts into an AI service unless your organization's policy permits it.

Common mistakes

Assuming silence means agreement

The prompt cannot infer consensus from a transcript unless agreement is explicit.

Guessing relative dates

Without the meeting date and timezone, ‘next Friday’ may be normalized incorrectly.

Publishing without review

Names, commitments, and sensitive statements should always be checked by a responsible attendee.

Questions, answered

Frequently asked questions.

Can I use a full meeting transcript?

Yes, within the model's input limit and your organization's privacy policy. For very long transcripts, process logical sections and preserve links or timestamps back to the source.

How does the prompt avoid inventing action items?

It explicitly prohibits invented owners and dates and asks the model to mark missing fields. You should still verify the output against the notes.

Can this format work for client meetings?

Yes. Add client roles in the team context and review the follow-up for confidentiality, commitments, and tone before sending it externally.

What if the meeting notes are incomplete?

The template surfaces ambiguous and missing information in a confirmation section. Use that list to follow up with attendees rather than asking the model to fill gaps.

Not quite your use case?

Build the prompt your situation actually needs.

SecondBrain turns your goal, audience, context, constraints, and required output into a reusable prompt.