EN

Anthropic launches Claude Tag, an always-on AI agent built to live inside Slack!

Nada Salam

1- Anthropic has introduced Claude Tag, an AI agent that joins Slack as a persistent team member and builds memory of a company's workflows over time.
2- Organizations can create separate Claude instances for different departments, with strict data isolation between each.
3- The agent is currently in beta for Claude Team and Enterprise customers.

The latest:

Anthropic is bringing its AI into the daily flow of work. Claude Tag joins Slack not as a chatbot to ping on demand, but as a standing team member that reads assigned channels, retains context, and acts on it — without being summoned each time.

Details:

  • Claude Tag builds a running memory of company projects, naming conventions, and past decisions from every channel and connected tool it has access to. It can answer questions about old conversations, surface sales figures, or pull incident reports without manual searching.
  • Companies can spin up multiple Claude instances scoped to different teams. A sales-focused Claude won’t share memory with an engineering one, and neither can access the other’s data or tools, according to Anthropic.
  • An ambient mode lets the agent intervene proactively — flagging unresolved threads, surfacing relevant past discussions, or reminding teams of pending issues it detects in ongoing conversations.
  • Admins control which channels Claude can read and which tools it can connect to. Claude Tag can be added to private channels but cannot surface information from them externally.

What to watch:

The beta is limited to paid customers for now. Broader adoption will test how organizations manage an AI agent with persistent, accumulating memory — and how much institutional context they are comfortable letting it hold.

 

 

What to read next