Anthropic launched Claude Tag, an AI agent that joins Slack as a kind of team member. Once an administrator adds it to a workspace and grants access to chosen channels, anyone can type @Claude in a thread to hand it a task while they work on something else. Claude breaks the request into stages, works through them with the tools it has been given, and posts the result back to the thread.
The product is available in beta for Claude Enterprise and Team customers, runs on Claude Opus 4.8, and replaces Anthropic’s earlier Claude in Slack app, which it will retire on August 3 after a 30-day migration window.
What distinguishes Claude Tag from a standard chatbot is that it behaves more like a shared coworker than a personal assistant. A single Claude identity serves a whole channel, so everyone can see what it is doing and pick up a task someone else started. It also accumulates memory over time, learning a team’s vocabulary and decisions so users do not have to re-explain context, and it can draw on other channels and data sources when permitted, though Anthropic says it never reads from private channels.
An optional ambient mode lets it act on its own, flagging useful updates and following up on stalled threads, and it can work asynchronously, even scheduling tasks for itself over hours or days. Cat Wu, Anthropic’s head of product for Claude Code and Cowork, said the power lies in being able to “tag it the same way that you would a coworker.”
Anthropic points to its own use as evidence. The company says about 65% of its product team’s code is now created through an internal version of Claude Tag, and that staff also use it to chase product metrics, work support tickets and find the root cause of bugs.
Access is meant to be tightly controlled through an agent-identity model: administrators scope which tools, data and channels each instance can touch, and memories stay siloed, so a setup built for sales cannot pass knowledge to one built for engineering. Admins can cap token spend per organization and per channel and review a full log of every task and who requested it.
The Enterprise Play
Claude Tag is Anthropic’s most aggressive move yet to embed itself where work actually happens, rather than in a separate tab. It arrives as the company pushes hard into the enterprise market, valued at $965 billion and preparing for a public offering, and after one corporate-spending index showed Anthropic overtaking OpenAI in enterprise adoption for the first time.
The form factor also raises real questions for buyers. A single organization identity with broad channel access concentrates risk and puts more weight on governance. Ambient mode can become noise if enabled too widely, which is why Anthropic suggests leaving it off until teams understand its failure modes. Pricing is token-based and not fully detailed, and an always-on agent could consume very differently from a tool used on demand. Reliability matters more too, since an always-on teammate makes any outage costlier.
A Contested Channel
The launch lands in one of the most fought-over arenas in enterprise AI. Slack, owned by Salesforce, has been positioning itself as an agentic operating system and rolled out dozens of new Slackbot capabilities this year, while rivals race to plant agents inside the same channels.
Anthropic has been building toward this for months, integrating Claude with Slack in late 2025, launching interactive Claude apps in January and releasing Claude Managed Agents in April with early adopters like Notion and Asana. Slack’s general manager framed the shift as making AI multiplayer, moving it out of private chats into the open channel. Anthropic says it plans to extend Claude Tag well beyond Slack, to the many other places teams work.
