Claude can now join your Google Meet calls. Here's what it actually does

Most AI note-takers do one job: they sit in your call, listen, and send you a summary afterwards. Useful, but you still do all the work after the meeting ends.

This month, Anthropic tried something different. In November 2026, it launched Claude Meetings, a beta that lets Claude join your Google Meet calls as a visible participant. Not just to take notes, but to start the work while you're still talking: draft the plan, sketch a timeline on a whiteboard, look up a number, answer a question in the chat.

I read Anthropic's launch post and its help page. Here's what it actually does, who can use it, and what I'd watch out for.

Want to see it in action? Check out Anthropic's own announcement of Claude Meetings.

#What Claude does in the call

Once Claude joins, it shows up like any other participant and posts a notice in the chat, so nobody is surprised.

During the call, you can talk to it. Say something like "Claude, make a doc" or "Claude, make a whiteboard", and it starts building while the meeting goes on. Anthropic says it can make docs, whiteboards and slides.

Anthropic also launched Claude Whiteboards in beta the same week, and whiteboards work right inside Meetings. So a planning call can end with the board already filled in.

You can also ask it questions, out loud or in the Meet chat. One detail I liked: by default, Claude answers in writing in the chat. It doesn't interrupt the call with a robot voice. When you ask out loud, it checks who's speaking each time.

If you've connected work tools, Claude can use them too. Anthropic's own example is a budget question: Claude asks for permission to look it up, then posts the answer in the chat. It always uses a person's connected tools with that person's permission, asking the first time.

#What you get after the call

When the meeting ends, Claude creates a meeting page in claude.ai. It has:

• a summary and the action items
• the full transcript
• everything Claude built during the call

People from your organisation who were on the invite get a recap email with a link. Later, you can ask Claude about past meetings in a normal chat, like "what did we decide about the launch date last week?"

This is the part I find most interesting. The meeting doesn't end with "someone should write this up". The first draft already exists.

The same week, Anthropic also launched Claude Sheets in beta. It's the same idea in a spreadsheet: Claude works inside a shared sheet next to your team, so the numbers from a call can land somewhere useful.

#How you add it

There's no special invite address. You go to Meetings in Claude, connect Google Calendar, pick a meeting and click Add Claude. You can also set it to auto-join all your meetings, or all except one-on-ones.

Claude can stay for up to eight hours and warns you before it leaves. You can ask it to stop listening, or remove it from the meeting page at any time.

#Who can use it right now

Not everyone, yet:

• It's a beta for Team and Enterprise plans. On Team plans it's on by default, and on Enterprise an owner has to switch it on.
• It works with Google Meet only for now. Anthropic says other platforms are coming.
• It's not available for organisations with zero data retention, HIPAA-ready setups or customer-managed encryption keys.

On cost: during the beta, notes, transcripts and answers are included on Team and Enterprise. Answers that use connected tools count towards your normal usage. Docs, whiteboards and other things Claude makes in a meeting are free until 10 February 2027.

#The privacy side

An AI that sits in every meeting raises fair questions, and Anthropic has clearly thought about some of them.

• It doesn't save audio. It keeps a transcript, a summary and descriptions of what was on screen.
• Outside guests are controlled by admins. When a company keeps it internal-only, Claude stays out of calls with people from outside, and leaves if one joins partway through. That includes another company's note-taker bot.
• It announces itself in the chat every time.

Anthropic also says plainly that you should make sure everyone on the call is OK with being transcribed. I'd go one step further: tell people at the start, every time, even if the bot already did. It costs five seconds and saves an awkward conversation later.

#What I think

For small teams like the ones I work with, the hard part of a meeting is rarely the talking. It's the hour afterwards: writing the notes, turning "we should" into tasks, building the first version of the plan. That's exactly the hour Claude Meetings is trying to remove.

It's early, it's Google Meet only, and it's not for every company's privacy rules yet. But the idea is right. An AI in your meeting shouldn't just remember what you said. It should help you start on it.

#The bottom line

Claude Meetings turns Claude from a note-taker into a teammate that works during the call. If your team is already on a Team or Enterprise plan and uses Google Meet, try it on one internal planning call this week, and see how much of the follow-up is already done when you hang up.

#Sources

• Introducing Claude Meetings, now in beta — Claude
• Use Claude Meetings — Claude Help Center