Anthropic launches Claude Tag for Slack

Anthropic launched Claude Tag, a beta product for Claude Enterprise and Team customers that puts an AI agent into Slack as a channel member rather than a private assistant. Users summon it by tagging @Claude with a task in natural language; the agent breaks the request into stages and works through connected tools and data sources, replacing Anthropic’s earlier, simpler Claude-in-Slack app.

The distinguishing design choice was persistence and shared visibility: one Claude instance serves an entire channel rather than spinning up a fresh session per user, so, in Anthropic’s framing, “anyone can see what it’s working on, and can pick up the conversation from where the last person left off.” The agent builds up memory of channel history and connected data over time, aiming to cut down on colleagues re-explaining context, and can optionally act proactively — flagging relevant information or following up on stalled tasks unprompted. It runs on Claude Opus 4.8 and gives administrators control over which tools, channels and monthly token budgets it can use.

Anthropic said internal adoption was already substantial: 65% of its product team’s code was written through the company’s own internal version of the tool, with use also expanding into sales, support and debugging. The figure was self-reported and unaudited, but signalled how far the company was betting that Claude’s role inside workplace chat tools would shift from answering questions to executing multi-step work autonomously.