Anthropic turns Claude into a persistent Slack teammate

Unlike previous iterations that required direct prompts, Claude Tag maintains a continuous presence within a channel. It functions as a shared identity, enabling team members to pick up collaborative tasks exactly where a colleague left off. The system operates on a scoped permission model, where administrators dictate which specific channels and data sets the AI can access, ensuring that sensitive legal information remains cordoned off from engineering discussions.

Beyond basic task delegation, the tool features an ambient mode designed to monitor team progress. It can intervene when tasks stall or flag information gaps discovered across different areas of the company. This move mirrors a broader industry push toward context-aware agents, putting Anthropic in direct competition with Microsoft’s Copilot and intelligence layers from startups like Glean. For organizations, the shift represents a move toward AI that functions less like a search engine and more like a coworker with a long-term memory of company operations.