Enterprise

Anthropic launches Claude Tag, planting a persistent AI teammate inside enterprise Slack channels

The new beta replaces Anthropic's existing Slack app with an always-on, multiplayer Claude that learns a company's context, monitors channels, and works asynchronously — Anthropic's sharpest move yet into the collaboration layer.

Photo: Unsplash / Priscilla Du Preez — A team of coworkers collaborating around a laptop in an open office

Anthropic launched Claude Tag on Tuesday, a beta agent that lives inside Slack channels, learns a company’s context as it lurks, and works asynchronously when summoned with an @-mention. It replaces the Claude in Slack app the company shipped in October 2025, and reframes the product from a single-user chat helper into something Anthropic and Salesforce are openly positioning as a coworker.

The framing matters. The old app was, in Cat Wu’s words, “very single-player, whereas Claude Tag is built to be interactive and multiplayer.” Wu, Anthropic’s head of product for Claude Code and Cowork, is selling the same vision Slack GM Rob Seaman described as “making AI multiplayer.” Translation: the agent observes channels, retains persistent memory, gathers facts from elsewhere in the org when given permission, and exhibits what Fortune called an “ambient” behavior that lets it proactively update employees and chase forgotten threads. Admins keep controls over data governance, channel access, tool access, and token spend.

The competitive map is the real story. Microsoft has Graph, expressed through Copilot and Work IQ. Snowflake and Databricks are angling to be the back-end for tacit organizational knowledge. Glean is building the intelligence layer between model and enterprise data. Claude Tag is Anthropic’s bet that the collaboration surface itself, not the warehouse beneath it, is where the agent should live.

Anthropic offers an internal proof point that doubles as a flex: Claude Tag is already approving and incorporating 65% of the code changes its own product team submits. The beta is open to Claude Enterprise and Claude Team customers, with Claude Code wired in.

The market context sharpens the move. Anthropic is valued at $965 billion, now ahead of OpenAI, and Ramp’s May AI Index, covering more than 50,000 U.S. companies, showed 34.4% of firms paying for Anthropic versus 32.3% for OpenAI, the first time Anthropic has led on business adoption. Lotus 1-2-3 didn’t lose to a better spreadsheet; it lost to the operating system underneath it. Anthropic is trying to be the layer, not the app.

Sources