← Back to Videos
Cloud Architecture

Anthropic Launches Claude Tag — Multiplayer Claude Inside Slack

Anthropic just launched Claude Tag — a Slack-native beta where you @Claude into channels, delegate tasks async, and Claude works like a teammate with persistent context memory + to

📅 25 June 20269:50✍️ Rahul Kumar

Overview

Anthropic has launched Claude Tag — a Slack-native beta that turns Claude into a genuine team member inside your Slack workspace. Rather than switching between a chat interface and your work tools, you @mention Claude directly in channels, assign it tasks asynchronously, and it works alongside your team with persistent context and access to your tools and codebase.

The most striking validation: Anthropic's own product team reports that 65% of their code is now generated by their internal version of Claude Tag. This is not a demo — it is the workflow Anthropic's engineers use every day.

What Claude Tag Actually Does

Five core capabilities ship with Claude Tag:

  • Channel presence: @Claude can be mentioned in any channel it has been added to, making it accessible to the whole team rather than siloed to individuals
  • Async task delegation: Assign Claude a task in a thread — it works through it and returns results, similar to assigning work to a colleague
  • Persistent context memory: Claude remembers prior conversations, decisions made in channels, and project context across sessions
  • Tool and codebase access: Claude can be wired to your code repository, internal tools, and data sources
  • Team context awareness: Claude operates with awareness of the channel's history and the team's working patterns

The 3 Workflow Shifts

1. AI as a Team Resource, Not an Individual Tool

Most AI tools are personal — one person's Claude, one person's Copilot. Claude Tag is shared infrastructure. The whole team can see Claude's contributions, build on them, and correct them. This changes the accountability model: AI output becomes a team artefact rather than a private shortcut.

2. Team Context Flows Into AI Decisions

Because Claude operates inside the channel, it has access to the decisions, constraints, and discussions your team has already had. It does not need to be briefed on context that already exists in Slack — it can read it.

3. AI Takes Work Off the Queue

With async delegation, Claude Tag shifts from answering questions to completing work items. A developer can @Claude to write a test suite, a product manager can ask it to draft a spec, an ops engineer can have it investigate an alert — all without switching context.

How It Compares

vs ChatGPT Workspace integrations: ChatGPT's Slack integration is primarily a question-answering interface — it responds to prompts but does not operate as a persistent team member with memory across conversations.

vs Microsoft Copilot in Teams: Copilot in Teams is deeply integrated with the Microsoft 365 ecosystem. Claude Tag does not have that breadth yet, but offers stronger reasoning quality for complex technical and coding tasks.

Availability and Access

Claude Tag is available in beta for Anthropic Enterprise and Team plan customers. It requires Slack and is currently in a limited rollout. Teams interested in the beta can apply through Anthropic's enterprise sales team.

Key Takeaways

  • Claude Tag brings Claude into Slack as a persistent team member with memory and tool access
  • Anthropic's 65% internal code generation stat is the strongest possible proof-of-concept
  • Async delegation is the differentiator — this is about completing work, not just answering questions
  • The team-resource model changes how AI output is reviewed and owned — a governance consideration for enterprise deployments
  • Available on Enterprise and Team plans — worth evaluating for any Slack-heavy technical or product team

Watch on YouTube

▶ Watch Now

Opens in YouTube

Share on LinkedIn

One click — copies a ready-to-post update about this video

About the Author

Rahul Kumar is a Senior Cloud and AI Architect at Microsoft with 13+ years of enterprise experience across Azure, AWS, and GCP.

Book a Discussion