← Index12 entries · By Task
Section IV · For the Task

Prompts for summarizing meetings

Prompts that turn raw meeting notes into clean decisions, action items, and follow-ups.

§ Overview

The value of a meeting evaporates if nobody captures who agreed to do what, and that capture is tedious enough that it often doesn't happen. The Summarize Meeting Notes Into Action Items prompt fixes exactly that: paste your messy notes or a transcript and it returns a clean summary with decisions, named owners, and follow-ups, so the meeting actually produces a record someone can act on.

The surrounding prompts handle adjacent summarizing work. The Summarize a Long Document Into Key Points prompt condenses a sprawling doc into an executive summary and quotable stats, the Facilitate a Sprint Retrospective prompt structures the discussion and its action items, and the Summarize an Audit Trail for a Client Walkthrough prompt turns a dense log into something a client can follow. The Build a One-Page Travel Brief prompt is summarizing of a different flavor, compressing logistics into one scannable page. The error AI helps you dodge is the summary that records what was said but loses what was decided. Tell it to separate decisions from discussion and assign every action an owner, and you get notes people trust.

§ Field Notes

What makes a good prompt for summarizing meetings

A strong meeting-summary prompt tells the model what structure you want out: decisions, action items with owners and due dates, open questions, and a short narrative summary, in that order. It also helps to name the attendees so the model can attribute ownership correctly instead of leaving every follow-up unassigned.

Quality also depends on what you feed it. Raw transcripts work, but flag the parts that are tangents so they don't get equal weight with the decisions. Ask the model to explicitly mark anything it's inferring versus what was clearly stated, because a confidently invented action item is worse than a missing one when people are relying on the notes to know what they own.

§ Pro Tips

Get sharper results

  • 01Ask for a fixed structure (decisions, then action items with owner and date, then open questions) so every summary is scannable and nothing important hides in prose.
  • 02Name the attendees in the prompt so the model can attribute each action item to a real owner instead of leaving follow-ups orphaned.
  • 03Tell it to flag any action item it inferred rather than heard, so you can verify ownership before sending the notes to people who'll be held to them.
  • 04For recurring meetings, paste last week's summary too and ask it to track which prior action items were closed and which carried over.
§ FAQ

Common questions

Can I feed a raw transcript or do I need to clean it up first?

Raw transcripts work well, and cleaning isn't required. It helps to note where the off-topic stretches are so the model doesn't weight a tangent like a decision. The messier the input, the more valuable it is to ask for a clear decisions-versus-discussion split in the output.

How do I make sure action items get assigned to the right person?

List the attendees by name in the prompt and ask the model to attribute every action item to one of them, marking any it's unsure about. Then do a quick human pass, since the model can mishear who volunteered for what in a crowded discussion.

Is it safe to summarize confidential meetings with AI?

That depends entirely on your organization's tooling and data policy. Check whether your AI tool offers a no-training or enterprise mode before pasting sensitive material, and strip names or figures you're not cleared to share if you're unsure.

§ The Prompts · 12
№ 008productivity

Summarize Meeting Notes Into Action Items

Turn messy meeting notes into a clean summary with decisions, owners, and follow-ups.

For
chatgpt·claude
№ 042writing

Summarize a Long Document Into Key Points

Condense long documents into an executive summary, key bullets, and quotable stats.

For
claude·chatgpt·gemini
№ 054business

Facilitate a Sprint Retrospective

Plan a full sprint retrospective with agenda, prompts, and action-item templates.

For
chatgpt·claude
№ 077productivity

Script a Difficult Feedback Conversation

Script a difficult feedback conversation using the SBI model with a collaborative, direct tone.

For
chatgpt·claude
№ 112productivity

Build a One-Page Travel Brief for an Executive

Generate a scannable one-page executive travel brief with flights, meetings, and emergency contacts.

For
chatgpt·claude
№ 113business

Summarize an Audit Trail for a Client Walkthrough

Generate a client-friendly audit trail summary with timeline, flagged inconsistencies, and prep notes.

For
claude·chatgpt
№ 198business

Board Update Memo With Metrics, Asks, and Honest Lowlights

Produces a candid, scannable board update with a metrics table, honest lowlights, and specific asks.

For
chatgpt·claude
№ 200business

Summarize a Sprint Retro Into Themes and Owned Actions

Clusters messy retro notes into themes with root causes and a table of owned, dated actions.

For
chatgpt·claude
№ 203business

Executive Summary of a Long Document for Time-Poor Leaders

Condenses a long document into a layered executive brief tailored to one reader's priorities.

For
chatgpt·claude
№ 204productivity

Build a Tight, Time-Boxed Meeting Agenda

Generates a time-boxed meeting agenda with owners, outcomes, and async/cut recommendations.

For
chatgpt·claude
№ 207productivity

Turn Messy Notes into Clear Action Items

Extracts owners, due dates, decisions, and open questions from unstructured notes.

For
chatgpt·claude
№ 212productivity

Summarize a Daily Standup into a Shareable Update

Condenses raw standup notes into a scannable async update with blockers and at-risk items surfaced.

For
chatgpt·claude