Anthropic released Claude Tag, a Slack-based agent that transforms Claude into a shared team assistant for Enterprise and Team customers. The beta launched June 23 with plans to expand beyond Slack to other workplace platforms.

Claude Tag replaces the existing Claude in Slack app and runs on Opus 4.8. Administrators have until August 3, 2026 to migrate from the previous integration, with a 30-day opt-in period.

The agent operates differently from individual Claude usage. Teams can summon Claude by typing @claude in any authorized channel, where it reads thread context, breaks down requests into steps, and posts results back to the same conversation.

Claude Tag can continue work across multiple days, schedule follow-ups, and maintain channel-specific memory. When connected to development tools, it can monitor backlogs, prepare meeting briefings, pull metrics, and create draft pull requests from bug reports.

Enterprise-focused deployment

The feature uses an "agent identity" model where Claude operates through dedicated service accounts rather than individual user credentials. This enables channel-scoped permissions, separate memory between workspaces, and comprehensive audit logging.

Administrators control which channels Claude can access, what tools and repositories it can use, and spending limits per channel or organization. Channel work bills to the organization while direct messages bill to individual user accounts.

Anthropic is offering launch credits to eligible customers: $25,000 for Enterprise organizations and $2,500 for Team organizations with at least 10 paid seats. Credits expire September 1, 2026.

The company positions Claude Tag as an evolution of its Claude Code and Claude Cowork features into "multiplayer AI work." Anthropic reports that 65% of its internal product team's code comes from its own version of Claude Tag, making this both a customer release and a test of AI-native operations.

Claude Tag will expand to additional workplace platforms beyond Slack in coming months.