Claude Tag: the Anthropic AI that works like a teammate in Slack

Anthropic, the company behind the Claude assistant, introduced a new tool on June 23, 2026 that shifts how AI shows up at work. It is called Claude Tag, and the idea fits in a sentence: instead of chatting with a bot in a separate window, you talk to it directly in Slack, the way you would with a colleague.

The distinction sounds minor. It is not. Until now, AI of this kind ran mostly one-on-one: you ask a question, you get an answer, and nobody else sees the conversation. Claude Tag goes the other way. It lives in your team channels, follows the thread of a discussion and can take a task from start to finish while you move on to something else.

1. What is Claude Tag?

Claude Tag is what Anthropic describes as a virtual employee that works in Slack. In practice, a team member asks it to do something, and the bot breaks the request into several steps it carries out on its own. When it finishes, it posts the result straight into the channel where everyone can see it.

Cat Wu, product lead for Claude Code and Cowork at Anthropic, sums up the difference well: "Claude Code, Cowork and chat are very solo, whereas Claude Tag is designed to be interactive and multiplayer." It is no longer a private exchange between you and an AI. It is a tool that slots into collective work.

There is also a behaviour the company calls ambient: Claude Tag can follow up on its own, provide status updates and flag a forgotten task without being asked again. For a team juggling a thousand small things, that kind of quiet nudge can make a real difference.

2. A multiplayer AI that learns your company

The most interesting piece is the shared identity. The whole company uses the same Claude instance. That means one employee can start a request and a colleague can pick the work up where it was left, with nothing to re-explain. The bot keeps the thread.

Claude Tag also learns information specific to your organization over time, from one channel to the next. You do not have to repeat context on every request: who does what, how you name your projects, what your habits are. The more it is used, the better it knows the house.

To give a sense of what that looks like in practice, Anthropic mentions that within its own product team, 65 percent of Claude Tag’s code changes are approved and merged. That is an internal figure, to be taken with the usual caution, but it shows the tool is more than a demo: it does real work that reaches delivery.

• You hand it a task in Slack, in ordinary language.
• It breaks the task into steps and works through them one at a time.
• It delivers the result in the channel and can follow up when something stalls.
• The whole team shares the same AI and the same history.

3. What does this change for a small business?

For a small or mid-sized company, the appeal is easy to see. Many of the tasks that eat time are not complicated, just repetitive: preparing meeting notes, gathering information scattered across several channels, writing a first draft, following up with an internal client. Those are exactly the things a teammate built into your messaging can take on.

The other advantage is that it stays where your people already work. If your team lives in Slack all day, you are not adding yet another tool to learn. You talk to Claude the way you talk to a colleague. For a small business without a full IT department, that simplicity counts as much as the capability.

That said, connecting an AI to your internal conversations is not a casual move. You need to decide which channels it can see, which data it can touch and who is allowed to hand it sensitive work. That kind of governance is precisely what an IT partner can help you put in place. At OKTO Solutions, it is part of what our managed IT services cover, and we can discuss it whenever suits you if you write through the contact page.

4. Security and control: what Anthropic put in place

Anthropic clearly thought about the data question, because an AI moving through your company channels raises fair concerns. Access to sensitive information and tools can be limited very precisely. Administrators choose, channel by channel, which tools, which information and which memories Claude can use.

There are also spending caps, applied both per person and across the organization. That prevents unpleasant surprises and keeps the invoice under control, a point that reassures people when a tool works in the background.

For a company, the right instinct stays the same as with any new technology: start small, on one or two channels, with clear access rules, then widen once the team is comfortable. A gradual, well governed rollout beats a broad and fuzzy one.

5. A trial version, not the finished product

Worth keeping in mind: Claude Tag ships first as a trial, what Anthropic calls a research preview. For now it is offered in Slack to customers on the Claude Enterprise and Claude Team plans, with wider availability planned later. If you are not on one of those plans, you cannot use it yet.

It also means the tool will move. Features will change, some will be added, others will disappear. That is normal for a product at this stage. If the idea of an AI teammate appeals to you, now is a good time to watch how it evolves before folding it into your daily processes.

Frequently asked questions

What is Anthropic’s Claude Tag?

It is a tool launched on June 23, 2026 that puts the Claude AI directly into Slack. You hand it a task, it breaks the work into steps, completes it on its own and delivers the result in the conversation, much like a virtual employee shared by the whole team.

Who can use Claude Tag right now?

At launch, Claude Tag is offered as a trial (research preview) in Slack, limited to customers on the Claude Enterprise and Claude Team plans. Anthropic plans to widen access afterward.

Is Claude Tag safe for company data?

Anthropic lets administrators limit, channel by channel, the tools, information and memories Claude can use, with spending caps per person and for the whole organization. As with any tool connected to your data, clear governance from the start remains essential.

Want to bring in AI without losing control of your data?

Tools like Claude Tag show how quickly AI is arriving inside the software your employees already use. To get value from it without exposing your information, a clear plan is worth having. Have a look at our IT services for small business or write through the contact page to talk it over with our team in Trois-Rivières.