What A Time To Be Alive
Choose language: English

Claude Meetings lets AI join video calls

Anthropic is testing Claude Meetings with business customers. The AI assistant joins a video call as a visible participant, takes notes, answers questions and drafts a plan or whiteboard while the meeting is still going on.

Title card with the words Claude in white and Meetings in coral orange against a starry sky, with a dark planet with an orange glow in the bottom right and the Anthropic logo in the top right.
Claude Meetings lets Anthropic's AI assistant take part in video calls.

Anthropic, the company behind the AI assistant Claude, announced Claude Meetings . With it, Claude takes part in a video call as an ordinary participant that everyone can see in the list. The feature is in a test phase, a beta, for companies with a Team or Enterprise subscription. For now it only works with Google Meet; other platforms will follow later.

What Claude does in the meeting

Participants can ask Claude a question by saying "Claude" out loud or by typing in the meeting chat. By default, Claude answers in that chat, not with a voice. Say "Claude, make a document of this plan", and Claude starts on it straight away. It can also make a whiteboard or a presentation. Colleagues who were invited to the meeting can already work on that document during the discussion.

Afterwards, everything is on a separate page for the meeting: the notes, the transcript of the conversation, the action points and whatever Claude made. Anyone who later wants to know what was decided can ask Claude about it there. Claude keeps a transcript of the conversation, a summary and descriptions of what was shown on screen, but no audio recording.

Claude can also use the work apps that someone has connected to Claude, for example to look up a figure. This only happens with that person's permission. If someone asks for this out loud, Claude checks each time, for now, who was speaking.

Privacy and costs

Claude announces itself in the meeting chat. Anthropic asks users to check in advance whether everyone agrees to the conversation being transcribed. An administrator decides whether Claude may join calls with people from outside the company. With a Team subscription this is allowed by default; with Enterprise, an administrator has to allow it first. Companies with extra strict rules on storing data cannot use the feature yet during the test phase.

Notes, the transcript and ordinary answers are free during the test phase. Work with connected apps counts towards the subscription's usage. Documents and whiteboards made in a meeting are free until 10 February 2027.

With Claude Meetings, Anthropic joins the race to become the AI assistant in the office. OpenAI released its own work tools earlier this autumn, and Microsoft and Google are also building AI into their office software.

What does this mean for the road to the singularity?

An AI that summarises a meeting is already common. What is new is that Claude gets to work during the meeting itself: while people are talking, the plan is already on paper and everyone can improve it right away. The step from talking to doing, which now often takes days because someone has to write up the notes afterwards, becomes much shorter.

For Anthropic's competitors, that is a signal. Whoever supplies the assistant that sits in the meeting also sees what a company is working on, and so becomes the place where the work starts. Microsoft, Google and OpenAI will therefore have to follow. If that trend continues, an AI assistant will more and more often join as a colleague who takes on tasks, not as a tool you switch on afterwards. That only works if participants can see what is being stored and companies decide what the assistant may do. Anthropic is already building that in, with a notice in every meeting and settings for administrators.

What is the technological singularity? Read the explanation →

Sources

  1. Claude
  2. TechCrunch
  3. support.claude.com