Skip to content
PromptifyLab

Business & Operations

Meetings & Notes prompts

8 prompts Free · no sign-up Works in ChatGPT, Claude & Gemini Search & filter these

Transcripts into decisions, actions and owners — without the invented commitments. Below are 8 copy-ready prompts. Fill in the [BRACKETS], copy, and paste into ChatGPT, Claude, Gemini or any capable assistant.

A meeting summary is only useful if someone who was not there can act on it. Most AI summaries fail that test: they recount the conversation instead of extracting what changed.

The 8 prompts

Beginner 4 blanks to fill

Turn raw meeting notes into decisions, actions and owners

Convert messy notes or a transcript into something people can act on.

Prompt
Turn these meeting notes into a structured summary.

MEETING: [MEETING NAME]
DATE: [DATE]
ATTENDEES: [NAMES AND ROLES]

NOTES/TRANSCRIPT:
"""
[PASTE]
"""

Produce exactly these sections:

1. DECISIONS MADE - only things actually decided. For each: the decision, who made it, and what it supersedes if anything. If nothing was decided, write 'No decisions were made.' Do not pad this section.

2. ACTION ITEMS - a table: Action | Owner (a named person) | Due date | Blocked by. If no owner was named, write OWNER UNASSIGNED in capitals. If no date was given, write DATE UNASSIGNED.

3. OPEN QUESTIONS - things raised and not resolved, with who raised them.

4. DISCUSSED, NO ACTION - topics that came up and went nowhere. One line each.

Rules:
- Never invent an owner or a date. Unassigned is a finding, not a gap to fill.
- Do not editorialise or summarise tone.
- If someone disagreed and was overruled, record it in Decisions.

What you get: Four clean sections with unassigned owners and dates flagged in capitals rather than invented.

Tip: The 'never invent an owner' rule is the whole value. Default model behaviour is to assign the most plausible person, which creates fake accountability.

Open in Written for Claude, ChatGPT, Gemini · Reviewed September 18, 2026
Beginner 4 blanks to fill

Write an agenda that makes a meeting shorter

Design an agenda around decisions rather than topics.

Prompt
Write a meeting agenda designed to be as short as possible.

MEETING PURPOSE: [WHAT NEEDS TO HAPPEN]
ATTENDEES: [NAMES AND ROLES]
TIME AVAILABLE: [MINUTES]
BACKGROUND: [CONTEXT, OR PASTE PRIOR NOTES]

Structure each agenda item as:
- Item title
- Type: DECISION / INPUT NEEDED / INFORMATION
- Decision to be made (one sentence, phrased as a question with a yes/no or pick-one answer)
- Who must be present for this item
- Minutes allocated
- Pre-read required (yes/no, and what)

Rules:
- Anything typed INFORMATION should not be in a meeting. For each one, say what it could be instead (email, doc, async comment) and remove it from the time budget.
- Total allocated minutes must be at most 80% of the time available.
- Order items so that anyone only needed for one item can leave early - put those first.
- End with 'Who can skip this meeting entirely and why'.

What you get: An agenda with a time budget, explicit decisions, and a list of people who should not be invited.

Tip: The 80% rule and the 'who can skip' line are what actually shorten meetings. Most agendas fail by budgeting 100% of the time to content and none to discussion.

Open in Written for Claude, ChatGPT, Gemini · Reviewed September 18, 2026
Intermediate 5 blanks to fill

Prepare for a difficult meeting

Anticipate objections and prepare responses before you are in the room.

Prompt
Help me prepare for a difficult meeting. Play devil's advocate against my position.

MY POSITION: [WHAT I WANT TO HAPPEN]
WHY I WANT IT: [MY REASONING]
WHO I AM MEETING: [ROLE AND WHAT THEY CARE ABOUT]
THEIR LIKELY POSITION: [WHAT I THINK THEY WANT]
CONSTRAINTS I CANNOT CHANGE: [BUDGET, TIMELINE, POLICY ETC]

Produce:

1. THE STRONGEST CASE AGAINST ME - written as they would argue it, at full strength. Do not soften it.
2. MY THREE WEAKEST POINTS - where my reasoning is thin, and what evidence would fix each
3. LIKELY OBJECTIONS - a table: Objection | Why they will raise it | My response in two sentences | What I should NOT say
4. THE ZONE OF AGREEMENT - what we probably both want, stated in their language not mine
5. MY FALLBACK LADDER - four positions from ideal to walk-away, with the trigger for moving down each rung
6. ONE QUESTION to open with that would make them explain their real constraint

Do not reassure me. If my position is weak, say so in point 2.

What you get: A full adversarial brief including your own weak points and a structured fallback ladder.

Tip: 'Do not reassure me' matters. Without it the model validates your plan, which is worse than useless before a negotiation.

Open in Written for Claude, ChatGPT, Gemini · Reviewed September 18, 2026
Beginner 4 blanks to fill

Summarise a long meeting for someone who missed it

Produce a catch-up brief calibrated to how much the reader needs to know.

Prompt
Write a catch-up summary for someone who missed this meeting.

WHO MISSED IT: [NAME AND ROLE]
WHAT THEY NEED IT FOR: [e.g. they own one of the action items / they are the approver / general awareness]

TRANSCRIPT OR NOTES:
"""
[PASTE]
"""

Write in three layers so they can stop reading when they have enough:

LAYER 1 - THIRTY SECONDS (max 60 words): what changed, what they now need to do.

LAYER 2 - THREE MINUTES (max 300 words): the decisions and their reasoning, anything that affects this person's work, anything they would have objected to had they been there.

LAYER 3 - FULL DETAIL: everything else, organised by topic, with who said what where it matters.

At the very top, before Layer 1, put a single line: 'YOU NEED TO: [the one thing]' or 'YOU NEED TO: nothing'.

What you get: A layered brief with a one-line action header, readable in 30 seconds or 10 minutes.

Tip: Layering beats summarising. Different readers need different depths, and one summary always fails half of them.

Open in Written for Claude, ChatGPT, Gemini · Reviewed September 18, 2026
Beginner 4 blanks to fill

Chase overdue action items without being annoying

Draft follow-ups that get a response.

Prompt
Draft follow-up messages for overdue action items.

OVERDUE ITEMS:
[PASTE LIST: action, owner, original due date, days overdue]

MY RELATIONSHIP TO THESE PEOPLE: [peer / their manager / their internal client / external]
CHANNEL: [email / Slack / Teams]
HOW URGENT IT ACTUALLY IS: [blocking others / important not urgent / nice to have]

For each person write one message that:
- Opens with the specific item, not a pleasantry
- States what it is blocking, factually, without guilt-tripping
- Offers a concrete out: a new date, a smaller version of the task, or reassignment
- Ends with a question that can be answered in one line
- Is under 80 words

Match the register to the relationship and channel. Do not use 'just circling back', 'gentle reminder', 'friendly nudge', 'as per my last message', or any exclamation marks.

If an item is genuinely nice-to-have and weeks overdue, say so and recommend dropping it instead of chasing it.

What you get: One short, specific message per person, plus a recommendation to drop anything not worth chasing.

Tip: Banning the stock phrases is what makes these read as human. Those five phrases are the tell that a follow-up is automated.

Open in Written for Claude, ChatGPT, Gemini · Reviewed September 18, 2026
Advanced 2 blanks to fill

Extract commitments from a transcript and check them against the last meeting

Find out what was promised last time and whether it happened.

Prompt
Compare two meeting records and report on follow-through.

PREVIOUS MEETING:
"""
[PASTE PREVIOUS NOTES OR TRANSCRIPT]
"""

CURRENT MEETING:
"""
[PASTE CURRENT NOTES OR TRANSCRIPT]
"""

Produce a table with columns: Commitment (from previous) | Owner | Status in current meeting | Evidence.

Status must be one of: DONE (explicitly reported complete), IN PROGRESS (mentioned, not complete), NOT MENTIONED, ABANDONED (explicitly dropped), or CHANGED (scope or owner moved).

Evidence must be a short quote from the current transcript, or 'no mention'.

Then list:
- Commitments that silently disappeared (NOT MENTIONED) - these are the ones to raise
- New commitments made in the current meeting that duplicate an old one
- Anyone who has now missed two or more commitments

Be literal. Do not infer that something was done because it was not raised as a problem.

What you get: A follow-through table with quoted evidence, plus a list of commitments that quietly vanished.

Tip: 'Do not infer that something was done because it was not raised' is doing real work here. Silence is the most common failure mode and models read it as success.

Open in Written for Claude, ChatGPT, Gemini · Reviewed September 18, 2026
Advanced 5 blanks to fill

Run a blameless post-mortem

Structure a review of something that went wrong without turning it into a search for a culprit.

Prompt
Facilitate a blameless post-mortem.

WHAT HAPPENED: [INCIDENT OR FAILURE]
IMPACT: [WHO WAS AFFECTED AND HOW]
TIMELINE: [PASTE WHAT YOU KNOW, ROUGH IS FINE]
WHO WAS INVOLVED: [ROLES, NOT NAMES]

Produce:

1. TIMELINE - reconstructed, with times where known and [TIME UNKNOWN] where not. Separate 'what happened' from 'when we found out'.

2. CONTRIBUTING FACTORS - grouped as: process, tooling, information, and workload/time pressure. Every factor must be a system condition, not a person's choice. Rewrite any 'X forgot to' as 'the process relied on memory with no check'.

3. WHAT WENT RIGHT - genuinely. What limited the damage.

4. THE FIVE WHYS - run it once, but stop when you reach a system condition rather than a person.

5. ACTIONS - a table: Action | Type (prevent / detect faster / reduce impact) | Owner role | Effort (S/M/L). Include at least one 'detect faster' action - most post-mortems only produce prevention.

6. WHAT WE ARE ACCEPTING - risks we are consciously choosing not to fix, and why.

Never name an individual as a cause. If the input names someone, replace with their role.

What you get: A structured blameless post-mortem with system-level causes, balanced actions, and explicitly accepted risks.

Tip: Section 6 is the one nobody writes. Naming what you are choosing not to fix stops the same action reappearing in every post-mortem for a year.

Open in Written for Claude, ChatGPT, Gemini · Reviewed September 18, 2026
Beginner 4 blanks to fill

Coordinate schedules and meeting requests for a busy manager

Handle calendar conflicts, reschedules and meeting requests with clear, polite messages.

Prompt
Help me manage my manager's calendar requests.

MY MANAGER'S PRIORITIES THIS WEEK: [PRIORITIES]
CURRENT CALENDAR (paste or describe): [MEETINGS, TIMES, TIME ZONES]
NEW REQUESTS: [LIST: who, what, how long, how urgent]
RULES MY MANAGER PREFERS: [e.g. no meetings before 10am, focus time on Friday, 15 min gaps]

Do this:
1. Identify conflicts and overloaded days
2. Suggest a plan: what to accept, move, shorten, delegate or decline, with a one-line reason each
3. Write the messages: reschedule requests, polite declines, and confirmations - short and professional
4. Write a 5-line daily briefing for my manager for tomorrow
5. Flag anything that needs my manager's decision rather than mine

Keep time zones explicit in every proposed time.

What you get: A conflict review, a prioritised accept/move/decline plan, ready-to-send messages and a daily briefing.

Tip: Always write time zones in proposed times. Most scheduling mistakes in international teams are time zone mistakes.

Open in Written for ChatGPT, Claude, Gemini · Reviewed September 25, 2026

Where AI actually helps here

  • Pulling decisions and actions out of an hour of talking
  • Writing the recap email in a tone the recipient will actually read
  • Separating what was decided from what was merely discussed at length

Where it falls down

  • Assigning actions to whoever was speaking nearby rather than whoever accepted them
  • Treating a long unresolved discussion as a decision because it sounded conclusive
  • Mangling names, products and numbers that the transcription already got wrong

The mistake almost everyone makes: Not giving it the attendee list

Paste the names and roles before the transcript. Without them you get “the team will follow up” instead of “Raj will send the mapping spec by Friday”, and you get actions attributed to whoever happened to be talking. Add one line: never attribute an action to someone who did not accept it out loud.

Free tool: Meeting Notes Prompt Builder

Runs in your browser. No sign-up, nothing uploaded.

Open the Meeting Notes Prompt Builder →

Questions people ask


What is the best prompt for meeting notes?

One that names the meeting type, the attendees with roles, who reads the output, and the exact shape you want back. A generic “summarise this meeting” produces a recap; those four additions produce something actionable.


How do I handle a transcript too long for the context window?

Split it into parts and label them: “part 2 of 4, do not summarise until I say so”. Have it extract decisions and actions from each part, then summarise from those extracts rather than from the raw text.