Anthropic Launches Claude Tag: Your Always-On AI Teammate for Slack

Anthropic today announced the research preview of Claude Tag, a new feature that deeply integrates AI into enterprise collaboration. By using @Claude in Slack channels or direct messages, users can access this always-on AI teammate at any time to get real-time insights, execute tasks, or continue conversations. The feature is available to Claude Enterprise and Claude Team customers, marking a shift from on-demand assistant to a permanent team member.

Unlike one-off Q&A sessions, Claude Tag maintains a unified Claude identity within a channel—all members see the same conversation history and can seamlessly pick up where others left off. Admins can fine-tune which tools, information, and channels Claude can access. For example, a Claude configured for a legal team won't be called into an engineering channel, isolating context and memory. Additionally, Claude Code integration can automatically route programming tasks in channels to a full development session on the web, with updates synced back to the discussion thread.

While this design boosts collaboration efficiency, it also raises the bar for enterprise data governance. Given that Slack often contains sensitive business information, strict permission controls and visibility boundaries will be critical for organizations deciding on large-scale deployment. As the research preview progresses, balancing AI's proactive learning with security and compliance remains a key challenge for Anthropic to validate.