Anthropic just dropped Claude Tag, an AI that lives inside Slack and learns your team’s work as it happens. Mention @Claude in a chat and it answers, remembers, even takes on tasks of its own. For the first time, an AI in the workplace is built to stay, not to visit.
The Gist
- Claude Tag is an always-on Claude that lives in Slack and remembers each channel’s history
- Anyone in the channel shares the same Claude, which can split a task into steps and post the result back
- The agent is in beta, only for Claude Enterprise and Claude Team customers for now
What Claude Tag is, in plain English
Slack is the chat app most teams use at work. It looks like a giant group conversation, split into channels for each project or team. Claude Tag is an AI assistant that sits inside those channels all day. You write a message, type @Claude, and the bot answers right there in the thread.
The interesting part is the memory. Most chatbots forget you the moment the window closes. Claude Tag does the opposite. It watches the channel it lives in, learns the names, the projects, the running jokes, and uses all of that the next time someone asks it for help.
Everyone in a given channel works with the same Claude. It is a shared identity, not a personal bot. If a colleague asked the bot to dig into a report yesterday, you can pick up the same conversation today and Claude already knows the context. The handover between teammates feels natural.
Admins keep control of the boundaries. They decide which channels each Claude can see, and which tools it can use. A Claude built for a legal team cannot suddenly look into an engineering channel. The cloistering protects sensitive conversations from leaking sideways.
When you give Claude a task, it does not stop at advice. It breaks the request into smaller steps, picks the tool it needs, and posts the work back into the Slack thread. Everyone in the channel sees the result, ready to react.
Why this is a big deal for daily work
The real twist is something called ambient mode. With it on, Claude does not wait for a mention. The agent joins the chat by itself to share an update, surface a piece of information from another channel, or remind people about a task left untouched for days. It behaves like a teammate who never logs off.
If you have ever forgotten to follow up on a Slack thread, you can already picture the change. Claude can hold the thread for you, raise it again when it matters, and connect a stray question to a project you discussed two weeks earlier. Memory is the feature, the chat is just the surface.
It also shifts how teams hand work off. Today, a manager often re-explains a project to a new joiner because the context lives in different people’s heads. A Claude that has been listening for months can give a clean summary on demand, without forcing anyone to dig through old messages.
Anthropic is not the only company chasing this idea. Microsoft has Copilot inside Teams, with similar promises around context and memory. Glean does the cross-tool version, looking through every app a company uses to find an answer. Snowflake and Databricks offer the database angle on the same problem. As we covered with AI loops, agents that work in the background are becoming the new normal.
For now, Claude Tag is in beta. Only Claude Enterprise and Claude Team customers can try it, which means most users will hear about it before they touch it. The launch is a test bed for Anthropic to see how a Claude with memory behaves in real teams before opening it more widely.
Keep learning on AI Noobies:
- AI Loops: How Bots Code While You Sleep
- iOS 27 AI Features: 8 Things Your iPhone Will Do
- Signal Boss: AI Chatbots Are Not Your Friends
How Claude Tag Will Show Up in Your Slack
Short term, you will probably hear about Claude Tag from someone at work before you ever see it yourself. If your company uses Slack and pays for Claude, expect a pilot inside one team first, often product, engineering or customer support. Other teams watch, then ask for their own Claude.
Medium term, the question of who reads what changes. Today, your manager reads what you write in Slack, and that is roughly the limit. With a Claude listening in the same channels, your messages also feed a memory that may serve answers for months. The line between a private thread and a company asset gets blurred.
There is upside too. If you have ever felt that you spend half your day repeating context to new colleagues or chasing forgotten tasks, an AI like this is built exactly for that pain. As Signal recently warned, an AI chatbot is not your friend, but in a work setting, a memory assistant can be a serious time saver.
The questions to ask are simple. Who decides what Claude can see in your team. How long does it keep memories. What happens when an admin opens access to another channel for a quick answer. These are not technical details, they shape what an employee can say in chat without it ending up in an AI’s long-term store.
The simplest takeaway is to know the word. Claude Tag is the new wave of AI at work, less assistant on a screen, more colleague in the chat. Knowing how it works puts you on the right side of the conversation when the rollout reaches your team.
Follow the story on AI Noobies.
About the Author
