Anthropic's Claude Tag Joins Slack as a Teammate, Writing 65% of Code
Anthropic's Claude Tag turns Slack into a multiplayer AI workspace where @Claude joins as a persistent team member, not just a chatbot
Takeaways
- Claude Tag launches in beta on Slack for Enterprise and Team plans, letting teams tag
@Claudeto delegate multi-step tasks asynchronously. - 65% of Anthropic's product team code now comes from their internal version of Claude Tag, signaling real internal adoption beyond a demo.
- Multiplayer by design: one shared Claude per channel builds persistent memory, so teammates can pick up tasks where others left off without re-explaining context.
- New "agent identity" access model gives Claude its own service accounts per channel, replacing per-user permissions with admin-scoped, auditable access.
- Ambient mode lets Claude proactively flag relevant updates and follow up on stalled threads without being explicitly tagged.
- Replaces the existing Claude in Slack app; admins have 30 days to migrate, with a launch credit issued to eligible orgs. Read the announcement.
Anthropic just shipped Claude Tag, a product that fundamentally changes what it means to have AI in your team's workflow. Rather than opening a chat window and typing at a model, you tag @Claude in a Slack channel the same way you'd loop in a colleague -- and it gets to work.
More than a Slack bot
The distinction matters. Claude Tag isn't a chatbot sitting in a sidebar. Claude joins as a team member, with access to the channels and tools you choose. Tag Claude with a request, and it breaks the task into stages, works through them with the tools it has access to, and responds in the thread with what it creates. That means writing or merging pull requests, running data analysis, or helping resolve an incident -- all without anyone switching context.
Three properties make this different from prior Slack AI integrations:
- Multiplayer by design. Within a given Slack channel, there's one Claude that interacts with everyone. Anyone can see what it's working on, and can pick up the conversation from where the last person left off.
- Persistent memory. As Claude follows along with its channel, it builds more context about the work. Users don't need to explain things to it from scratch over and over again.
- Ambient initiative. If "ambient" behavior is enabled, Claude will proactively keep you updated about whatever it thinks you might need to know. It'll flag relevant information from across the channels it's in and the tools it's connected to, and follow up on threads or tasks that have gone quiet without being resolved.
The number that tells the whole story
Tagging @Claude is now one of the main ways Anthropic gets things done internally. Today, 65% of their product team's code is created by their internal version of Claude Tag. That's not a benchmark number -- it's a live production stat from the company that built the thing. The same pattern is spreading well beyond engineering: teams are tagging Claude to chase down product metrics and data, work through support tickets, and help find the root cause of tricky bugs.
For context, Claude Code crossed $2.5B in annualized revenue by February 2026 , and Claude Tag is explicitly positioned as its evolution. The trajectory here is steep.
A new access model under the hood
The multiplayer setup creates a real security challenge: if multiple people are steering Claude in a shared channel, whose permissions apply? Anthropic's answer is a new concept called agent identity -- Claude gets its own accounts in each system it touches, set up by admins and tied to the workspace rather than any individual user.
In a channel where Claude Tag is active, Claude isn't acting on behalf of a single user. It has its own account in each system it touches: it posts in Slack as the Claude app, opens pull requests as the Claude GitHub App, and queries your warehouse under a service account provisioned by an admin.
Admins define what Claude can access at the workspace level, then override it per channel. The engineering channel gets GitHub and the data warehouse; the sales channel gets the CRM. Memory and access respect those boundaries: what Claude learns in a private channel never appears in the wider workspace. Every action is logged under Claude's own service accounts, making audits straightforward.
Admins also control spending directly. Administrators can set limits for token spend both for the organization and for individual channels, and can view a log of everything that @Claude has done, along with who requested each task.
Async as a first-class feature
One underrated capability: Claude Tag is built to work while you're not watching. Set Claude a task, and you can focus on your other priorities while it works. It can also schedule tasks for itself, pursuing a project autonomously over hours or days. This is a meaningful shift from the current generation of AI coding tools, which mostly require a human in the loop to keep things moving.
Claude Code is described as the most capable autonomous agent, particularly for async and Slack-based workflows. You can assign a task to Claude Code directly from a Slack message and come back to a finished PR. For distributed teams that live in Slack, this changes how work gets done.
Where it fits in the competitive landscape
The agentic coding space is crowded. GitHub Copilot, Cursor, and OpenAI Codex are all pushing into autonomous territory. But Claude Tag's Slack-native, multiplayer design is a distinct bet. The Slack integration arguably exposes a weakness in competing AI coding assistants -- they're still trapped in isolated environments. While GitHub Copilot offers Slack integration for basic code snippets, and Cursor provides thread-based debugging, none seem to match Claude Code's full workflow automation directly within team conversations.
The philosophical split is real: Copilot lives in your editor and completes your code as you write. Claude Code lives in your terminal and executes complete tasks autonomously. Claude Tag pushes that autonomy further into the social layer of work -- the place where decisions actually get made.
That said, the tools aren't mutually exclusive. In practice, many senior developers use both: Copilot in their IDE for writing as they go, Claude Code in a terminal for structured work sessions and automation. The two tools are complementary more than competitors.
What it means beyond engineering
The most interesting signal in the announcement isn't the 65% code stat -- it's the mention that the pattern is spreading beyond engineering. Support tickets, product metrics, incident response: these are workflows that live in Slack but have historically required humans to bridge the gap between conversation and action. Claude Tag is a direct attempt to close that gap.
Slack has long been the connective tissue between teams within an organization -- and now, it's the connective tissue between humans and AI agents, enabling them to work together right where key decisions are made. Anthropic is betting that whichever AI tool owns that layer owns a disproportionate share of enterprise value.
How to get started
Claude Tag is available in beta for Claude Enterprise and Team customers. It replaces the existing Claude in Slack app, and admins have 30 days to migrate. Anthropic is issuing a launch credit to eligible organizations so teams can try it without immediately burning through their token budget. The product runs on Opus 4.8.
Setup takes four steps:
- Pair Claude Tag with your Slack workspace
- Connect Claude to your tools (GitHub, data warehouse, etc.)
- Set a monthly token spend limit for your org
- Test in a private channel before rolling out broadly
Expansion to other platforms beyond Slack is planned, though no timeline has been given. For teams already deep in the Anthropic ecosystem, the official docs and the agent identity access model explainer are the right places to start before provisioning access at scale.