Anthropic has unveiled Claude Tag, a new way for teams to collaborate with its AI assistant directly inside Slack, positioning the model as a persistent digital colleague rather than a one-off chatbot. The feature, announced on June 23, 2026, is available immediately in beta for Claude Enterprise and Team customers.

The launch marks a notable evolution of Anthropic's Claude Code and Cowork products, pushing the assistant from a tool you summon for a single task into something that lives alongside a team and takes on work over extended periods. For more context on this story, see our ongoing latest AI developments.

How Claude Tag Works

With Claude Tag, Claude joins a Slack channel as if it were a human team member. Administrators grant the assistant access to selected channels and connect it to the tools, datasets, and even codebases the team uses day to day. Once installed, anyone in the channel can simply type `@Claude` to delegate a task and then move on to other work while the assistant handles it.

Claude builds context by remembering relevant information from the channels it inhabits, and it can plan out tasks to complete later. When it finishes, it reports back in a Slack thread with whatever it has produced.

According to Anthropic's announcement, tagging `@Claude` has already become one of the primary ways the company gets work done internally. The company disclosed that 65% of its product team's code is now created by its internal version of Claude Tag, and the practice is spreading well beyond engineering to chasing down product metrics, working through support tickets, and tracing the root cause of bugs.

Four Behaviors That Set It Apart

Anthropic emphasized several capabilities that distinguish Claude Tag from a conventional chat interface:

Multiplayer Collaboration

Within a given Slack channel there is a single, shared Claude that interacts with everyone. Any teammate can see what the assistant is working on and pick up the conversation from where the last person left off. Anthropic describes this as closer to interacting with a collaborative teammate than firing off isolated prompts.

Learning Over Time

As Claude follows the conversation in its channel, it accumulates context about the team's work. Users do not need to re-explain the same background every time. If granted permission, Claude can also learn automatically from other Slack channels and connected data sources, though Anthropic noted that it does not read or report from private channels.

Taking Initiative

When an optional "ambient" mode is enabled, Claude proactively surfaces information it thinks the team might need, flags relevant updates from across its channels and connected tools, and follows up on threads or tasks that have gone quiet without resolution.

Asynchronous Execution

Users can assign Claude a task and turn their attention elsewhere while it works. The assistant can also schedule tasks for itself, pursuing a project autonomously over hours or even days. Anthropic said its own staff increasingly delegate tasks to many instances of Claude in parallel, allowing humans to focus on oversight and decision-making.

Why Slack, and What Comes Next

Anthropic chose Slack as the launch surface because it is a natural home for collaborative work between teams and AI, and where much of Anthropic's own day-to-day operations already happen. The company signaled that its goal is to expand Claude Tag more broadly so that teams can tag `@Claude` in the many other places they work beyond Slack.

Users can also send Claude direct messages, in which case it responds privately using the personal tools and connectors they have configured.

The Bigger Picture

Claude Tag arrives as the AI industry races to build so-called agentic systems, assistants that can act with greater autonomy rather than merely answering questions. Competitors including OpenAI and Google have similarly pushed toward agents that can browse, execute code, and complete multi-step workflows. Anthropic's Slack-native approach, however, leans heavily on the workplace messaging tools where knowledge workers already spend their hours, lowering the barrier to adoption.

The emphasis on persistent memory and proactive behavior also reflects a broader industry shift. Rather than treating each interaction as stateless, vendors are building assistants that retain context across sessions, learn an organization's conventions, and operate as long-running background workers.

For Anthropic, the internal productivity claims are striking. If roughly two-thirds of a major product team's code is now machine-generated and reviewed, it offers a concrete data point in the ongoing debate over how much software engineering work AI can absorb — and how quickly that share is growing.

Availability

Claude Tag is rolling out today in beta to Claude Enterprise and Claude Team customers on Slack. Anthropic has not yet announced pricing details specific to the feature or a timeline for broader availability beyond Slack, but the company indicated that expanding where Claude can be tagged is a priority.

For organizations already invested in the Claude ecosystem, the launch effectively turns the assistant into infrastructure that is always present in the channels where decisions get made — waiting to be tagged in, and increasingly, volunteering to help before being asked.

---

Stay Ahead of AI

Get the latest AI news, analysis, and breakthroughs — all in one place.

Read more AI news →