Based on the Lead with AI AI Leader Advanced program.
Most leaders know their meetings cost too much. The bigger cost often starts after the call, when someone has to write up what was decided and chase who does what.
AI meeting assistants capture the conversation and draft the summary, decisions and action items for you. Your job shifts from taking notes to staying present and confirming what matters.
What does an AI meeting assistant do?
Most tools follow the same pattern. They capture the audio, turn it into a timestamped transcript with speakers, and use a language model to draft the summary, decisions and action items. Many also draft a follow-up message or send tasks to your project tracker.
Should you use a dedicated note taker or a built-in copilot?
A dedicated note taker, such as Granola, works across Zoom, Teams, Google Meet and any other call on your computer. A built-in copilot lives in the platform you already use, where your IT team has most likely settled the data questions. If your company runs on Zoom, its own AI Companion covers similar ground.
The five meeting assistants at a glance
| Tool | Works with | How it listens | Plans |
|---|---|---|---|
| Copilot in Teams | Microsoft Teams | Built into the meeting | Microsoft 365 Copilot license |
| Gemini in Google Meet | Google Meet and in-person meetings | Built into the meeting | Workspace Business and Enterprise, Google AI Pro and Ultra |
| Granola | Any call on your Mac or Windows computer, plus in person on iPhone | Listens on your device, with no bot | Free plan, Business from about $14 per user per month |
| ChatGPT Record and Meetings | Any call or conversation on your Mac | Records on your Mac, with no bot | Record on Plus and up, Meetings beta on Pro and Business |
| Claude Meetings | Google Meet | Joins as a visible participant | Team and Enterprise, in beta |
Which meeting assistant should you use?
Start with the platform your meetings already run on.
- Microsoft Teams: Copilot in Teams helps during the call and turns the recap into actions.
- Google Meet: Gemini in Google Meet writes notes and an optional transcript into Google Docs.
- Several platforms: Granola takes notes on any call from your computer, with no bot in the meeting.
- Calls outside the usual platforms: ChatGPT Record and Meetings record on your Mac, without a bot.
- Building while you talk: Claude Meetings drafts documents and whiteboards during Google Meet calls.
How should your meeting routine change?
- During the meeting, stay present. You are a contributor, not the scribe.
- After the meeting, confirm what matters. Check the summary, clarify anything ambiguous, confirm owners and dates, then send the follow-up.
- Use the open questions as the starting point for the next agenda.
The payoff is not just fewer notes. It is fewer "what did we decide?" messages, faster handoffs for people who missed the call and a record you can search weeks later.
A CODO prompt for the follow-up
Paste any transcript into ChatGPT, Claude, Gemini or Copilot with this prompt, built with our CODO framework. Replace the bracketed parts.
Character: You are an experienced chief of staff who turns meeting transcripts into clear follow-up. Objective: I am the [role] at [company]. Turn the attached transcript of [meeting name] into a post-meeting package I can act on today, for [the people who attended]. Do: - Separate real decisions from things that only sounded like agreement. - Use speaker names from the transcript for action items, and flag any item without a clear owner. - Give every action item an owner, a next step and a deadline, and flag dates nobody actually stated. - Keep it skimmable, with short headers and bullets. Don't: - Don't invent anything that is not in the transcript. If the transcript is thin, say so. - Don't bury open questions. They are often the most valuable part. Output: A post-meeting package with five sections: a 3 to 4 sentence summary, decisions made, action items, open questions, and a follow-up message I can review and send myself.
What should you keep in mind?
- Treat the output as a draft, not a record. Meeting AI is strong at structure and weak at nuance, implied commitments and exact names or numbers.
- Get consent. Tell people when AI is taking notes, and follow your company's policy and local recording rules.
- Keep sensitive meetings out. Legal, HR and board discussions may need no recording at all.
- Review commitments before they go out. Check legal, financial and performance points against the transcript.
- Send follow-ups yourself. Let AI draft them, and send them only after you review them.
Want a worked example? How Jakob Knutzen (GM, Miro) uses AI before, during and after every meeting shows a leader's full routine, and Your meeting architecture is broken helps you decide which meetings deserve to exist.
Key takeaways
- Start with the platform your meetings already run on.
- Stay present during the call and confirm what matters afterward.
- Treat AI notes as a draft, and check names, numbers and commitments.
- Tell people when AI is listening.
- Send the follow-up yourself.
This overview is based on the Meeting Assistants module in Lead with AI's AI Leader Advanced program. Lead with AI helps senior leaders build practical AI fluency through hands-on courses and a membership community.
