Anthropic brings Claude Tag into Slack as an always-on AI teammate
Anthropic's new Slack-native agent joins channels as a permanent team member, builds context over time and can work on its own, marking the company's deepest push into the workplace and sharpening the race to embed AI inside the tools teams already use.
Anthropic has launched Claude Tag, a Slack-native AI agent that behaves less like a chatbot and more like a permanent member of the team. Announced on 23 June 2026 and available in beta for Claude Enterprise and Team customers, Claude Tag lets anyone in a Slack channel summon the assistant by typing @Claude and hand it a task to work on independently. The launch matters because it pushes AI out of private one-on-one chats and into the shared channels where teams actually plan, decide and get work done.
The product replaces Anthropic's earlier Claude in Slack app and marks the company's deepest move yet into the workplace. It also lands in fiercely contested territory. Salesforce, which owns Slack, has been positioning the platform as an operating system for AI agents, while rival labs race to plant their own assistants inside the tools employees already use every day.
What Claude Tag actually does
Built on Claude Opus 4.8, Anthropic's most capable publicly available model, Claude Tag is designed to be "multiplayer." Every member of a channel works with the same Claude identity, and the agent retains context over time, so teams do not have to brief it from scratch for each task. It can break a request into steps, use connected tools and data, and post results back into the thread. An optional "ambient" mode lets it act without being tagged, flagging stalled threads or surfacing updates on its own. Anthropic says about 65% of its own product team's code now runs through an internal version of the tool. Administrators can cap monthly token spend at the organisation and channel level, a nod to enterprise worries about runaway AI bills.
Rob Seaman, general manager of Slack, said the integration is about making AI "multiplayer," shifting it from a private exchange into the open channel. Cat Wu, who leads product for Claude Code at Anthropic, said the underlying capabilities already existed in some form, but the real change is in how employees interact with the agent. The company, led by co-founder and chief executive Dario Amodei, says it plans to bring Claude Tag to other platforms in the coming weeks. Rivals are moving too, with OpenAI rolling out workspace agents earlier this year and well-funded startups chasing the same prize.
How is Claude Tag different from a regular chatbot
A standard chatbot answers one person at a time and forgets the conversation once the window closes. Claude Tag is built to stick around.
Because it lives inside a channel, it shares a single memory with the whole team and keeps learning the group's vocabulary, decisions and ongoing work. It can also keep working asynchronously, picking up a task one person started and carrying it forward over hours or days while everyone else moves on.
For Indian enterprises, global capability centres and startups that run much of their daily work on Slack, Claude Tag points to where workplace AI is heading: agents that sit inside team conversations rather than waiting in a separate tab. As Anthropic extends the feature to more platforms, the question for adopters will be how much autonomy, and how much spend, they are willing to hand over.