On June 23, 2026, Anthropic introduced Claude Tag, a new way for teams to work with its AI: instead of opening a separate chat, you tag @Claude within Slack as if it were just another teammate and delegate tasks to it while you keep doing your own thing. It’s not an assistant you consult and then close: it’s a model that stays in your channels, remembers the context, and moves work forward autonomously. For any SMB that already collaborates in Slack or Teams, Claude Tag marks a trend worth understanding well before jumping on board —or not.
At Tisa we’ve been helping companies leverage technology with judgment, not by fashion, since 1987. So let’s explain exactly what this feature does, what it means for your business, and where the fine print is.
What Claude Tag is and why it matters
According to Anthropic’s official announcement
The difference from a traditional chatbot is the working model. Anthropic sums it up in four traits
It’s "multiplayer"
Within a channel there’s a single Claude that interacts with everyone. Anyone can see what it’s working on and can pick up the conversation where another teammate left off. It feels less like an individual chat and more like having a colleague on the team.
It learns over time
As it follows the channel’s conversation, Claude builds context about the work. That means, according to Anthropic
It takes initiative
If the "ambient" behavior is enabled, Claude proactively flags what it thinks you need to know: it points out relevant information from connected channels and tools, and follows up on threads or tasks that have been left unresolved
It works asynchronously
You assign it something and you carry on with your priorities while it moves forward. It can even schedule tasks for itself and pursue a project over hours or days. Anthropic claims that internally they already delegate tasks to many "Claudes" in parallel, and that 65% of their product team’s code is generated by its internal version of Claude Tag
From answering questions to executing tasks
The conceptual leap that Claude Tag represents is the same one we’ve been telling our clients about: AI is moving from answering to doing. A classic assistant returns text to you; an agent like this one breaks the assignment into phases, executes them using the tools it has access to, and replies to you in a thread with what it has produced.
Anthropic gives usage examples that have overwhelmed engineering
- Support and customer care: tag the AI to draft a response to a recurring ticket.
- Operations tracking: have it monitor a thread of orders or incidents and alert you when something gets stuck.
- Business data: ask it to gather and summarize metrics that today someone looks up by hand.
It’s exactly the territory of intelligent automation: combining task execution with the judgment that previously required a person.
The fine print: control, permissions, and cost
This is where an SMB has to think like a company, not like an enthusiast. Anthropic has designed Claude Tag with access control front and center
- Administrators define which tools and which information the model sees and in which channels. It’s like creating separate identities: a Claude configured for sales doesn’t share memory with one for engineering.
- You can set token spending limits per organization and per channel.
- A log is kept of everything
@Claudehas done and of who requested each task.
This is no minor detail: an AI that "learns from your channels" and "takes initiative" is powerful, but it requires deciding sensibly what data it accesses, who can delegate tasks to it, and how much it can spend. The feature is available in beta for Claude Enterprise and Team customers and works with the Opus 4.8 model
What your company should do about this
The fact that a tool exists doesn’t mean it’s for you, nor that it’s for tomorrow. But Claude Tag confirms a clear market direction: AI agents are getting inside the tools where you already work. If your team lives in Slack, Teams, or your ERP, the question stops being "do we use AI?" and becomes "which specific tasks make sense to delegate, with what controls and at what cost?".
And that’s exactly the work we do at Tisa from our custom development line: helping you identify measurable use cases, connecting them to your systems (Business Central, Power Platform, your own tools), and setting it up with the governance and permissions your business needs. It’s not about having AI for the sake of it, but about it solving something real without opening up a security hole or a hole in the budget.
Let’s talk about your case
If you’re interested in exploring how to apply AI agents like Claude Tag —or alternatives within the Microsoft ecosystem you already use— to a specific process in your company, we’ll help you assess it with no commitment. Write to us at info@grupotisa.com, call us at (+34) 971 305 885, or stop by grupotisa.com. We’ll study your case and tell you, honestly, what’s realistic today and what isn’t.