AI meeting notes and minutes: stop writing them yourself

The short answer

AI meeting notes work like this: the meeting is recorded and turned into text, and AI writes the summary, decisions and action points. Teams, Google Meet and Zoom can do it on most paid plans, tools like Fireflies and Otter cost about USD 10-20 a month, and the cheapest way is pasting a transcript into ChatGPT or Claude.

What AI meeting notes are

The meeting is recorded and turned into text. Then AI writes a short summary: what was discussed, what was decided, and who does what by when. You get it by email or in the meeting chat a few minutes after you hang up.

It works for video meetings and, with a phone or a small microphone on the table, for meetings in a room.

Option 1: Use what you already pay for

Check this first. You may have it already.

• Microsoft Teams: meeting summaries come with Microsoft 365 Copilot or Teams Premium. Without those, Teams can still record and transcribe, and you can paste the transcript into ChatGPT or Claude yourself.
• Google Meet: "Take notes for me" writes notes into a Google Doc. It's part of the Gemini features in most paid Google Workspace plans.
• Zoom: the AI Companion summary is included in paid Zoom plans at no extra cost. Switch it on in the settings.

Option 2: A separate note-taker

Tools like Fireflies, Otter and tl;dv join your meetings as a "bot" and send notes afterwards. They work across Teams, Meet and Zoom, can search old meetings, and push tasks into other tools. Most have a free tier and paid plans of about USD 10-20 per person a month.

They're a good choice if you meet clients on whatever platform they use. Check where the recordings are stored and whether there's a data processing agreement before you sign up.

Option 3: Record, then ask ChatGPT or Claude

The cheap, flexible way. Record the meeting (most phones can), get a transcript from your meeting tool or a transcription app, and paste it into ChatGPT or Claude with instructions like these:

Save that as instructions in a project, so you don't have to type it every time. See Setting up Claude Projects.

Rules for recording people

This matters more than the tool.

• Tell everyone at the start that the meeting is recorded and why. Most tools show a notice, but say it out loud too.
• Give people a way to say no. If a client doesn't want to be recorded, write notes the old way.
• Don't keep recordings forever. Keep the written notes, delete the audio after a set time, like 30 days.
• Be careful with sensitive meetings, like staff reviews, health or legal matters. Don't record those with a third-party bot.

Recordings of people's voices are personal data under GDPR. More on that in Is it safe to put company data into ChatGPT?

Getting good notes

• Say decisions out loud. "So we agree: Anna sends the offer by Friday." The AI picks that up far better than a nod.
• Use names. "Can you do that?" is hard to assign. "Can you do that, Per?" isn't.
• Read the notes before you send them. AI sometimes mishears names, numbers and technical words. Two minutes of checking saves confusion later.

What it saves

If you have five internal meetings a week and spend 15 minutes writing notes after each, that's over five hours a month. Even if you still spend three minutes checking, most of that time comes back.

Questions people ask

What is the best AI for meeting notes?

The one built into the meeting tool you already pay for: Copilot in Teams, Gemini in Google Meet or AI Companion in Zoom. If you meet on many platforms, a separate tool like Fireflies or Otter is easier.

Is it legal to record meetings with AI?

Yes, if you tell everyone at the start, have a good reason, and handle the recording as personal data under GDPR. Give people a way to say no, and delete recordings after a set time.

Can ChatGPT take meeting notes?

Not by joining the meeting itself, but it writes good minutes from a transcript. Paste the text in with clear instructions, as shown above.

Going further

Notes are most useful when they don't stay in an inbox: action points straight into your task list, client meeting notes saved to the right customer in your CRM, and a weekly summary of all meetings. Linking those up is a typical AI workflow project.