Two days after previewing it alongside Sheets and Whiteboards, Anthropic has started rolling out Claude Meetings, and X's own summary is already lining it up against Otter, Fireflies, Zoom and Microsoft.

Anthropic started rolling out Claude Meetings on Thursday, putting Claude into Google Meet calls as a visible participant that starts on the work while people are still talking. The official Claude account posted the rollout on X at 12:16 p.m. PT, and Anthropic's own announcement says it's in beta on Team and Enterprise plans.

That answers the question we left open on Tuesday. When Anthropic came for Microsoft Office on the first day of its IPO roadshow, Meetings was part of the pitch but missing from the company's rollout post. It wasn't clear whether customers could use it yet. Now they can.

People noticed. The rollout post passed roughly 350,000 impressions, 11,600 likes and 774 quote posts in under an hour. Posts mentioning "Claude Meetings" on X went from 1 in the 11 a.m. hour to 135 in the noon hour.

Claude takes its seat

Claude doesn't just sit in and listen. Teams invite it into a Google Meet call like any other colleague, and it shows up in the grid as a labeled participant. People can ask it questions out loud or in the chat, and it answers in writing.

The real pitch is what Claude builds while the meeting is still going. TechCrunch reports that it can draft the plan the team is discussing, map out a timeline on a Claude Whiteboard and pull numbers like last year's figures from connected work apps. When the brainstorm wraps, a whiteboard can be turned into a doc or a slide deck.

Each meeting gets its own page in claude.ai with the notes, the transcript and everything Claude made, according to Unite.AI. Invited colleagues can edit those outputs, and attendees can come back later and ask Claude about the meeting.

Anthropic has been eating its own cooking. As we reported Tuesday, Nicholas Lin, who leads Anthropic's knowledge work products, said Claude has sat in on nearly all of the company's meetings for about two months, mostly waiting for instructions.

The fine print on a bot that hears everything

An AI in every meeting raises an obvious question about who's listening, and Anthropic has put in a few guardrails. Claude posts a notice in the chat when it joins. It stays out of calls with external guests by default and leaves if one shows up, though admins can change that. It keeps a transcript, a summary and descriptions of what's on screen, but not the audio.

Connected tools run on each person's own permissions, but anything Claude posts in chat is visible to the whole call. On Team plans the feature is on by default. On Enterprise, admins have to switch it on. Anthropic also tells customers to get everyone's consent to transcription before adding Claude.

There are limits. It's Google Meet only, with other platforms "coming soon." It isn't available to organizations using zero data retention, a HIPAA-ready setup or customer-managed encryption keys, which rules out some of the most cautious buyers. Notes, transcripts and in-call answers are free on Team and Enterprise. Docs and whiteboards made during calls are free through Feb. 10, 2027.

The notetakers are in the crosshairs

The line going around X is that this is a lab shipping a whole category as a feature. X's own news summary of the launch puts Claude Meetings head to head with Otter, Fireflies, Zoom AI Companion and Microsoft's Copilot in Teams.

The comparison is a little off, and that's the point. Most meeting bots record, transcribe and summarize, so the work starts after everyone hangs up. Anthropic is selling the opposite: the plan, the timeline and the numbers get built during the call, in the same workspace where the team keeps working on them.

There's a twist. Anthropic says it's working with Gong, Attention, Granola, Fireflies.ai and Otter so users can send meeting transcripts to those tools, per Unite.AI. That isn't part of the beta yet. For now, Anthropic is plugging into the notetakers rather than trying to bury them.

The bigger fight is still with Microsoft and Google. They own the calendar and the meeting platform, and Zoom and Teams users can't use Claude Meetings at all yet.

What it means for founders building on the labs

If your product is something the model can do on its own, the lab can ship it as a feature inside a product your customers already pay for. Claude Meetings is the latest reminder, and it lands two days after Anthropic put a $1.3 trillion price tag on its IPO. A company with public-market investors to impress needs more than API revenue, and the applications its customers already use are the easiest place to find it.

The question for founders is no longer whether the lab will compete with them. It's what they own that the lab doesn't. Gong's answer is sales analytics built on years of call data. Granola's is a notes app people choose to open. A bot that only transcribes and summarizes has a much harder answer.

Our take

For years the notetakers have been getting companies used to having a bot on the call. Anthropic is collecting the payoff, with a bot that does the work instead of just writing it down.

It's a beta, it's Google Meet only and the most security-conscious buyers can't turn it on yet. None of that changes the direction. The labs are moving out of the chat window and into the meetings, docs and boards where work actually happens. Any startup that lives there should already be working out what it can defend.

Tuesday, Claude got a seat at the table. Thursday, it started taking meetings.