What Is Claude Tag? Anthropic's Slack AI Coworker, Explained

In late June 2026, Anthropic launched Claude Tag — and replaced its older "Claude in Slack" app with something much more ambitious: an AI teammate you tag into your channels that works on its own. If you've seen the launch coverage and want the plain-English version of what it actually does, who can use it, and what teams are debating, this is it.

What Claude Tag actually is

Claude Tag lets your team collaborate with Claude directly inside Slack. You @Claude into a channel — or DM it, or open it from the AI assistant side panel — and hand it work. According to Anthropic, Claude then:

• Works under your organization's identity, not a personal account.
• Builds context over time, remembering relevant information from the channels it's in across days.
• Follows up on its own — it can pick a task back up and check in proactively rather than waiting to be asked each time.
• Uses the tools and repositories an admin connects, at three permission scopes: organization-wide, workspace, and private channel.

The mental-model shift Anthropic is pushing is "multiplayer." Most AI assistants are a private 1:1 copilot. Claude Tag's work happens in shared channels the whole team can watch and hand off — one teammate starts something, another picks it up, and Claude carries the thread. That's the genuinely new idea, and it's the part early reviewers respect most.

Who can use it today

This is where expectations need a reality check. At launch, Claude Tag is:

• Slack-only. There is no Microsoft Teams version yet.
• In beta, on Claude Team and Enterprise plans — not free or Pro.
• Admin-gated. Only a workspace Primary Owner or Owner can set it up: provisioning Claude's org identity, connecting tools, and choosing which channels are in scope.

So the attention is broad — it was a loud launch — but hands-on access is narrow. If you're on a personal plan, or your company runs on Teams, you can't simply switch it on.

The open questions teams are raising

The launch earned real respect for the paradigm. The doubts cluster in one place: control and access. These are the questions worth asking of any autonomous workplace agent, Claude Tag included:

• Does it act before a human approves? An "ambient," always-on agent that can interject and follow up unprompted is powerful — and, for regulated or cautious teams, the thing they most want a gate in front of.
• One shared identity, broad reach. A single org-scoped Claude that any channel member can ask about everything it can see is convenient, but it cuts against least-privilege instincts and raises "which instance did that, and why?" audit questions.
• What's the data scope? It reads channel history and uses connected tools; teams want clarity on how long context is retained and exactly what it can reach.

None of these are disqualifying — they're the normal governance questions a brand-new autonomous product has to answer over time. Treat the capability claims as Anthropic's own until independent reviews land.

How this compares to an AI coworker for Slack and Teams

Junior is in the same category — an AI coworker that lives in your team chat and acts across your tools — so the honest comparison is about shape, not the underlying model (Junior routes work to frontier models, Claude among them). Where the two differ today:

	Claude Tag	Junior
Chat platforms	Slack only (beta)	Slack and Microsoft Teams
Availability	Team / Enterprise beta, admin-provisioned	Generally available, self-serve
Control model	Ambient / proactive; approval gates debated	Approval-gated by default — it drafts and asks before acting, and actions are audited
Focus	General-purpose teammate	Pre-configured for sales, marketing, and ops roles
What it's used for	Tasks + follow-ups in Slack	CRM hygiene, weekly reports, daily briefings, follow-ups across 3,000+ connected tools

The short version: if your team is Slack-first and wants a general autonomous teammate, Claude Tag is a serious new option. If your team lives in Microsoft Teams, or you want every outbound action approved before it happens and available today without an enterprise beta, that's the gap a coworker like Junior is built for — frontier-model intelligence shaped for where your team actually works.

For the full head-to-head, see Junior vs Claude Tag — and for the earlier Slack-assistant comparison Claude Tag replaces, Junior vs Claude in Slack. For the broader field, the best AI agents for Slack in 2026 walks through assistants vs coworkers vs workflow runners.

---

The bottom line: Claude Tag is a real step — Anthropic taking the "AI as a teammate, in the tools you already use" idea seriously, with a genuinely fresh multiplayer model. It's also early: Slack-only, beta-gated, and still answering the control-and-oversight questions that come with any agent that acts on its own. Whether you adopt it, wait, or run a Teams-native coworker alongside it, the category it's validating — an AI employee that works where your team works — is exactly the bet worth making.

If you want one you can hire today, in Slack or Teams, with approval gates on by default: meet Junior — free trial, $100 credit, first workflow live in minutes.