Anthropic Launches Claude Tag: @Claude Now Lives in Your Slack

Anthropic launched Claude Tag on June 23 in beta — a new product that embeds Claude directly into Slack as a persistent, always-on AI teammate for developer teams. Unlike the previous Claude in Slack integration that handled on-demand requests, Claude Tag joins channels as a shared team member, accumulates context across conversations over time, and can run multi-step coding and data tasks autonomously.

The distinction matters. Claude Tag is not a smarter chatbot inside Slack. It is an agentic system that connects to codebases, external data sources, and internal tooling — and remembers what it learns across sessions. Anthropic reports that 65% of its own product team's code is now generated through an internal version of Claude Tag.

What Claude Tag Does in Practice

Tag @Claude in any authorized channel with a request — write a feature, pull a product metric, triage a support ticket, investigate a bug root cause — and it breaks the task into stages, works through them using connected tools, and posts results back into the Slack thread. Teammates can see Claude's progress in real time and pick up a task that someone else started.

Three behaviors separate Claude Tag from prior integrations:

Multiplayer. One Claude instance serves the entire channel. Any team member can see what it is working on, and anyone can continue where the last person left off. This is closer to a shared contractor than a private assistant.

Persistent memory. Claude follows channel activity over time, building institutional context without requiring users to re-explain background. If granted permission, it can also learn from other channels and connected data sources across the organization.

Asynchronous execution. Tag Claude with a task and step away. It can pursue multi-step projects over hours or days, schedule its own follow-up tasks, and surface results without interrupting the team. Anthropic says it now spends more time delegating tasks to many Claudes in parallel than working through a single session.

An optional ambient mode takes this further. When enabled, Claude proactively monitors channels, flags relevant updates from connected tools, and follows up on unresolved threads without being asked.

Admin Controls and Security Model

Anthropic has built enterprise access controls into Claude Tag from the start. System administrators define which channels, tools, data sources, and memories each Claude identity can access. A Claude configured for engineering cannot leak its context into a channel set up for legal or HR. Admins set token spend limits at the organizational and per-channel level, and receive a full audit log showing every action Claude took and which user initiated each request.

Claude Tag runs on Claude Opus 4.8 and replaces the existing Claude in Slack app. Current customers have a 30-day window to migrate. Eligible Enterprise and Team organizations receive an introductory launch credit to trial the feature across their teams.

What It Signals

The race to embed AI agents inside the tools software teams already use is accelerating. OpenAI has enterprise integrations for ChatGPT and Codex. Microsoft's Copilot has structural depth across Teams and Microsoft 365. Claude Tag is Anthropic's answer to the same strategic question: where does AI take root in the real workflow, not the parallel tool that gets opened in a separate tab?

Getting @Claude into Slack threads — where millions of professional conversations happen daily — is a distribution bet as much as a product one. For developers, the practical question is whether a team-scoped ambient agent with persistent memory and codebase access changes how much they can delegate. Based on Anthropic's internal data, with 65% of product code already produced this way, the answer appears to be yes.

Claude Tag is available today in beta for Claude Enterprise and Team customers on Slack. Anthropic plans to expand to other platforms, with no timeline disclosed.