Anthropic Launches Claude Tag, an AI Teammate That Lives in Slack
Anthropic launched Claude Tag on June 23, 2026 — a new way to work with Claude that turns the assistant into a persistent teammate inside Slack. Instead of opening a separate app, anyone in a channel can tag @Claude to delegate a task, ask a question, or hand off an unfinished thread. Claude breaks the request into stages, does the work, and replies in the thread where the whole team can follow along. The feature is in beta for Claude Enterprise and Team customers and replaces Anthropic’s older “Claude in Slack” app.
General Audience
One Claude per channel, shared by the whole team
The headline idea behind Claude Tag is that it is multiplayer. Within a given Slack channel there is a single Claude that everyone interacts with, rather than a private chatbot session per person. As Anthropic puts it, “anyone can see what Claude has been working on, and can pick up the conversation from where the last person left off.” If a colleague asks Claude to draft a launch plan and goes offline, anyone else in the channel can step in, review the work in progress, and keep it moving.
Because it lives in the channel, Claude also builds context over time. Anthropic describes it plainly: “As Claude follows along with its channel, it learns ever more about the work.” The practical payoff is that teams stop re-explaining the same background — project history, naming conventions, who owns what — every time they ask for help.
Ambient mode, async work, and self-scheduling
Claude Tag can go beyond answering when called. With an optional “ambient” mode enabled, Claude proactively flags information it thinks you need from across the channels and tools it can reach, and it follows up on unresolved threads instead of letting them go stale.
It also works asynchronously. You can hand Claude a task and move on to something else while it runs in the background, and it can schedule work for itself — pursuing a longer project autonomously over hours or days. Anthropic frames this as an evolution of Claude Code: the same agentic capability that writes and reviews software, now made more proactive and built to work alongside a full team rather than a single developer. The current version runs on Anthropic’s flagship Opus 4.8 model.
Admin controls and what it means
Putting an autonomous agent inside a company’s primary chat tool raises obvious questions about access and oversight, and Anthropic has built guardrails for administrators. Admins control which tools, data, and channels each Claude can reach, and can create “separate Claude identities for different uses” — so, for example, a Claude scoped to legal channels cannot bleed context into engineering. Admins can also set token-spend limits at the organization and per-channel level, and audit every task with information about who requested it. Eligible organizations receive launch credits, and the old Slack app gets a 30-day migration window.
The most striking signal of intent is Anthropic’s own usage stat: the company says 65% of its product team’s code is now created by its internal version of Claude Tag. That positions the feature squarely in a broader enterprise race to give AI persistent organizational memory — alongside efforts from Microsoft, Snowflake, Databricks, and Glean — where the bet is that an assistant that quietly learns your company over time is far more useful than one you brief from scratch every session. For students and teams already living in Slack, it is a concrete preview of what an “AI coworker” actually looks like day to day.
Related Coverage
- Anthropic Releases Claude Opus 4.8 for Longer Agentic Coding — the flagship model that powers Claude Tag.
- Anthropic Ships Agent View: A Multi-Session Dashboard for Claude Code — an earlier step toward managing many parallel Claude sessions.
- Anthropic Plugs Claude Into Adobe, Blender, and Ableton With Nine New Connectors — Claude’s push to live inside the tools teams already use.
沪公网安备31011502017015号