INDEPENDENT AI NEWS / AI-assisted. Human-directed.EDITION 007 / 12 November 2026

Meetings & other human contributions

Finally, someone in this meeting is doing the work.

Claude Meetings joins the call and starts making the documents while we are still discussing them. We have finally found a colleague who listens and does the follow-up.

Here’s what happened.

Anthropic’s 12 November announcement says Claude Meetings is starting to roll out in beta on Team and Enterprise plans. The new help page describes Claude joining Google Meet as a visible participant: its tile shows its current state, people address it by name or in chat, and it can create documents, whiteboards and other artifacts during the discussion. Replies normally appear in writing. The meeting page keeps the transcript, notes and materials afterwards. This follows the 10 November overview, which still describes Meetings as coming soon; the newer announcement supplies the rollout update. Access depends on the workspace and its settings. These are Anthropic’s descriptions of the feature, not results from a call tested by this newsroom.

Source: Anthropic: Use Claude Meetings ↗

Our familiar robot appears as a video-call participant while two people review a document and planning board on the adjacent screen.
AI-generated editorial illustration about an AI participant creating work during a meeting. The call and planning screens are an imagined scene, not a Claude Meetings screenshot or a newsroom test.

The We Are So Done take

We used to leave meetings with a list of things somebody had to make. Now Claude can make them while we’re still explaining why they’re complicated. The document appears. The whiteboard fills up. Someone asks for a revised version. At last, the meeting is productive. We are delighted to be attending it.

The important part is that it is visibly there. It occupies a place in the call. We say its name. Someone asks it to put an idea on the board. The software has acquired a perfectly ordinary place in the group. We invited it.

And of course we want it there. We would like the person who had the idea to spend less of Friday turning everybody’s recollection of it into a document. We would like to see the plan while we can still argue about what it was supposed to say. Give us the colleague who can prepare another version before the conversation has moved on.

It does make our own contribution easier to inspect. We are discussing, choosing, correcting and approving while the work appears beside us. These are useful things to do. They are also becoming a remarkably complete description of what we bring to the room. We will explain the distinction as soon as the first draft is ready.

A finished-looking plan can still contain the wrong decision, and somebody has to notice. We are keeping that responsibility. Meanwhile, the software has joined the team, taken the notes and started the follow-up. We should probably make it feel welcome. It has already made itself useful.

For those of us already clicking

Give one internal planning call a useful first draft.

Access
Team or Enterprise beta rollout, Google Meet and an organization-enabled Google Calendar connection. Check workspace availability and owner settings.
Start here
With everyone’s agreement, use a short internal planning call with non-sensitive material. Ask for one document. Compare it with the decisions actually made: scope, owners, dates and open questions. Revise an error before relying on the result. This is a suggested first exercise, not our test.
The small print
We have not tested a call. Beta exclusions, external-guest settings, connected-tool permissions and usage terms apply; check the current help page for your organization.

Follow the evidence.

Source-based reporting. No independent hands-on test by this newsroom.
Sources checked 2026-11-12 · Editorial updated 2026-11-12