Claude Tag Turns Slack Into a Living AI Coworker That Never Forgets

News Summary

Anthropic launched Claude Tag on June 23, 2026 (Eastern Time), introducing a persistent AI teammate that lives directly inside Slack — moving beyond on-demand chatbot interactions toward a model that learns an organization's workflows, proactively surfaces insights, and completes long-running tasks autonomously. The product is now available in beta for Claude Enterprise and Team customers.

What Is Claude Tag?

Claude Tag transforms Claude from a reactive assistant into a continuous, channel-aware teammate. Powered by the Claude Opus 4.8 model, it joins designated Slack channels as a first-class member. Team members summon it with a simple @Claude mention, assign it tasks, and Claude breaks those tasks into stages — executing each one with the tools it has been granted access to, then reporting back in a threaded reply.

The defining shift from previous AI assistants is persistence. Rather than starting each conversation from scratch, Claude Tag builds and retains context by observing channel activity over time. The more it follows a channel, the richer its understanding of team jargon, recurring projects, stakeholders, and priorities becomes. Anthropic describes this as the model learning "ever more about the work" as it remains embedded in a team's daily communication stream.

Key Features and Capabilities

Multiplayer Collaboration

Unlike traditional AI chat interfaces that create one-on-one sessions, Claude Tag operates as a single shared identity within a channel. Every team member can see what Claude has been doing, pick up a conversation where a colleague left off, and delegate follow-up work without losing accumulated context. This makes asynchronous hand-offs seamless — a project manager can assign research in the morning and an engineer can review and extend Claude's findings in the afternoon.

Persistent Contextual Memory

Claude Tag maintains channel-scoped memory, meaning it remembers decisions, ongoing workstreams, and relevant data from the channels it monitors. Administrators can also grant cross-channel access so Claude can pull facts from related channels — for example, allowing a product channel Claude to draw on engineering status updates — while hard boundaries prevent sensitive knowledge from bleeding across unrelated teams.

Ambient Mode and Proactive Assistance

When ambient mode is enabled, Claude Tag does not wait to be tagged. It monitors activity across its connected channels and proactively surfaces relevant information: flagging items from other channels that a team should know about, following up on threads that appear to have stalled, and reminding teams of outstanding tasks. This turns Claude from a tool that answers questions into one that actively manages attention.

Asynchronous and Long-Horizon Work

Claude Tag can plan and schedule tasks that span hours or even days. A user can delegate a task in the afternoon, and Claude will continue working on it overnight — querying tools, synthesizing information, and delivering results when they are ready. This capability targets workflows where meaningful output requires extended processing time rather than instant replies.

Enterprise Security and Administrative Controls

Security was a central design consideration. Anthropic gives workspace administrators granular control over each Claude Tag instance:

• Scoped tool and data access: Admins define exactly which tools, data sources, and channels a given Claude instance can reach, preventing unauthorized access to sensitive information.
• Memory isolation: Each Claude Tag identity operates in its own memory silo. A legal team's Claude cannot seed memories into an engineering channel's Claude, keeping institutional knowledge appropriately compartmentalized.
• Token spend limits: Organizations can cap how much compute each Claude instance consumes, enabling predictable cost management.
• Activity logs: Full audit trails allow administrators to review what Claude has done, which data it accessed, and which actions it took.

The existing Claude in Slack application will be replaced by Claude Tag, with a 30-day migration window provided for current users to transition.

Anthropic's Own Usage and Industry Impact

Anthropic's announcement included a striking internal data point: 65% of the product team's code is now generated by an internal version of Claude Tag, which the company has used across engineering, product metrics analysis, support ticket handling, and debugging workflows. This suggests the company views embedded, channel-native AI as a material productivity lever rather than a supplementary tool.

Industry analysts note that Claude Tag positions Anthropic directly against enterprise AI initiatives from Microsoft (Copilot and Microsoft Graph), Snowflake, Databricks, and Glean — all of which are racing to build AI systems that capture and operationalize organizational knowledge. The key differentiator Anthropic is betting on is the depth of contextual understanding that comes from living inside team communication channels over extended periods, rather than querying documents on demand.

Availability and Pricing

Claude Tag is available in beta as of June 23, 2026 (Eastern Time) for organizations on Claude Enterprise and Claude Team plans. Eligible organizations receive introductory credits to get started. Anthropic has not published standalone pricing for Claude Tag at this time.