Anthropic Launches Claude Tag for Teams on Slack, Boosting AI Collaboration

Sixty-five percent of Anthropic's product team's code is already created by their internal version of Claude Tag, a figure that highlights the efficiency of this new tool. Anthropic has just introduced Claude Tag, an innovative way for teams to work directly with its Claude AI model, integrating it as a team member within Slack. This feature, launching in beta for Claude Enterprise and Team customers, aims to transform collaboration and productivity in the workplace.

Claude Tag allows users to tag @Claude in selected Slack channels, delegating tasks while the team focuses on other priorities. The AI can access tools, data, and even codebases granted by administrators. Once tagged, Claude builds context by remembering relevant information from the channels it's in and can plan out tasks to complete autonomously.

"Tagging @Claude is now one of the main ways we get things done at Anthropic."

This evolution of what was previously Claude Code makes the model more proactive and works better with a full team. The company, which already uses this tool extensively internally, is expanding its use beyond engineering to tasks such as tracking product metrics, managing support tickets, and identifying the root cause of tricky bugs.

One of Claude Tag's key advantages is its "multiplayer" nature. Within a given Slack channel, there is one Claude that interacts with everyone. This means anyone can see what the AI is working on and pick up the conversation from where the last person left off, fostering seamless, shared collaboration.

Furthermore, @Claude learns over time. As it follows along with its channel, it builds more context about the ongoing work, eliminating the need to explain things from scratch repeatedly. If granted the necessary permissions, Claude can even automatically learn from other Slack channels and data sources, developing crucial tacit knowledge to provide the best possible work.

Initiative is another standout feature. If "ambient" behavior is enabled, Claude will proactively keep the team updated about relevant information from across the channels it's in and the tools it's connected to. It will also follow up on threads or tasks that have gone quiet without being resolved, ensuring nothing falls through the cracks.

Claude Tag also excels in its asynchronous work capability. Users can set Claude a task and focus on other priorities, knowing the AI can schedule and pursue projects autonomously over hours or even days. At Anthropic, they have found this particularly helpful, delegating tasks to many Claude instances in parallel.

Security and control are fundamental to Claude Tag's design. System administrators can precisely specify which tools and information the model should have access to in each channel. This allows for the creation of separate "identities" for Claude, ensuring its memories and access remain scoped to the defined channels, preventing, for example, a sales Claude from accessing engineering data.

To get started, Claude Enterprise and Team customers in beta need to pair Claude Tag with their Slack workspace, give Claude access to necessary tools, set a monthly spend limit for the organization, and test it in a private channel. This new functionality replaces the existing "Claude in Slack" app, and administrators have 30 days to migrate. Anthropic is issuing an introductory launch credit to eligible organizations to facilitate the transition.