Related Articles
AI’s Next Big Workflow Idea Is Getting Loopy
AI News
Jun 23, 2026
Anthropic has released an early version of a new feature called Claude Tag. It places the company's AI assistant directly inside Slack, where it works less like a tool you open and more like a coworker who is always around. The feature is now available to Claude Enterprise and Claude Team customers. Anyone in a Slack channel can mention @Claude to ask a question, give it a job, or bring it into a conversation that is already underway.
At its simplest, Claude Tag is a permanent version of Claude that lives in your team's Slack channels. You do not have to switch to another app or open a separate chat. You just tag it, the same way you would tag a colleague.
This is not Anthropic's first move into Slack. People could already send Claude a direct message or tag it in a channel for quick help. A related tool, Claude Code in Slack, takes a coding request from a channel, runs a full coding session on the web, and posts the results back into the thread. Claude Tag goes further by adding two things those earlier tools lacked: lasting memory and a sense of continuity.
The key idea behind Claude Tag is that it stays and learns. As it follows along in a channel, it builds up an understanding of what the team is working on. If an administrator grants permission, it can also draw in useful information from other channels across the company. Over time, this turns a general-purpose chatbot into an assistant that knows your projects, your habits, and even your team's shorthand.
In a Claude Tag channel, the whole team talks to one shared Claude rather than each person using a private copy. That means everyone can see what the assistant has been doing, and anyone can pick up a task where a teammate left off. The work stays out in the open for the group instead of getting lost in private messages.
An assistant that remembers company information naturally raises questions about privacy and access. To address this, Anthropic gives administrators control over what each Claude can do. Admins decide which tools, data, and channels the assistant can reach, and each version of Claude stays limited to the channels it is assigned. For example, a Claude set up for the legal team cannot carry what it learns into an engineering channel.
When you assign Claude Tag a specific job, it breaks the work into steps and completes them using whatever tools it has access to. Once it is finished, it posts the result back into the Slack thread.
Claude Tag also has an ambient mode, in which it speaks up on its own without being prompted. In this mode it can share updates, point out useful information from around the company, and remind the team about tasks or conversations that have stalled. Anthropic describes the overall feel as working with a real colleague, one whose work is visible to everyone and who understands far more context than earlier tools did.
The headline benefit is productivity, but the bigger ambition runs deeper. Anthropic, like several of its rivals, is betting that the company that best captures an organization's knowledge, its decisions, its workflows, and its institutional memory will control the most valuable layer of enterprise AI.
The competition is already crowded. Microsoft is chasing the same goal through Graph, which powers Copilot and Work IQ. Snowflake and Databricks are pitching themselves as the back end that stores the data and knowledge AI agents rely on. Glean is building a layer of intelligence that understands company context and sits between the AI model and a company's data.
Seen in this light, Claude Tag is more than a handy Slack add-on. It is Anthropic's effort to weave itself into the daily flow of work and, one message at a time, learn how a company really operates.
Share your thoughts about this article.
Be the first to post a comment!
