Anthropic launches Claude Tag for Slack workplace workflows

Anthropic has launched Claude Tag for Slack, now in beta for Claude Enterprise and Team customers.

Claude Tag lets teams add Claude to selected Slack channels as a shared participant that can be tagged with tasks. Administrators can connect it to chosen tools, data sources and codebases, and set limits on access, channel scope and spending.

The launch expands Anthropic's effort to position Claude inside workplace software. Rather than keeping the model in one-to-one chats, Anthropic is pushing a version designed to sit inside group workflows, where multiple employees can see requests, follow progress and continue conversations in the same thread.

Anthropic says the system remembers relevant information from the channels it is allowed to access and can work through tasks in stages before replying in a Slack thread. It can also be configured to schedule work over longer periods, allowing staff to assign tasks and return later for the result.

Internal use

Claude Tag has already become a significant internal tool at Anthropic. The company says 65% of its product team's code is now created by its internal version of Claude Tag, and staff also use it to examine product metrics, review support tickets and investigate software bugs.

That figure offers one of the clearest indications yet of how heavily AI companies are beginning to use their own models in day-to-day work. It also suggests Anthropic sees collaborative use inside existing office software as a key route to wider adoption among business customers.

In Slack channels, there is one Claude instance for everyone in a given channel, making the interaction visible to the wider group rather than to a single user. Anthropic says this shared setup is intended to let colleagues pick up from previous exchanges without starting from scratch.

Another feature is what Anthropic calls ambient behaviour, which can be enabled to allow Claude to post updates without being directly prompted. This can include flagging information from connected tools and following up on unresolved threads or tasks.

Access controls

Anthropic has put strong emphasis on administrative controls in the launch. System administrators decide which channels Claude can join, which tools or information it can use, and how those permissions are separated across teams.

Memories remain limited to the channels defined by administrators. A setup created for sales, for example, would not pass memories to one configured for engineering, and engineering users would not gain access to sales data or tools through that arrangement.

Administrators can also review a log of Claude's actions and see who requested each task. They can set token spending limits for both an organisation and individual channels.

Slack shift

Claude Tag replaces Anthropic's existing Claude in Slack app, with administrators given a 30-day period to opt in to the migration. The product works with Opus 4.8.

Launching first in Slack reflects how central messaging platforms have become to internal coordination across technology companies and large businesses. By embedding the model in those channels, Anthropic is trying to make AI use less of a standalone activity and more of a routine part of team communication.

Anthropic frames Claude Tag as part of an evolution of Claude Code, extending the model from coding assistance to broader workplace delegation. That includes both engineering work and operational tasks that require access to company context spread across messages, tools and internal systems.

For customers, the appeal is likely to rest on whether the product can reduce repetitive coordination work without creating new risks around data access or inaccurate responses. Anthropic's answer so far is to keep deployment tightly controlled at the administrator level while allowing workers to interact with the model through simple Slack prompts.

The beta is limited to Claude Enterprise and Team users, signalling that Anthropic is initially targeting organisations that already have formal oversight of access, usage and budgets. Anthropic says 65% of its product team's code is created by its internal version of Claude Tag.