Anthropic launched Claude Tag on June 23, 2026, a research preview that positions Claude as a participant inside team communication channels rather than a tool invoked through a separate interface. The first integration is Slack, where Claude joins a workspace as a named team member.

The mechanics are straightforward: once added to a workspace, Claude can be mentioned directly in channels or threads. It can search and reference relevant Slack messages during a conversation, which means responses can be grounded in the actual context of that team's discussions — prior decisions, shared terminology, project history — rather than relying solely on what a user types into a prompt.

The distinction from existing Claude integrations matters here. Slack has had Claude connectivity before, allowing the model to pull in message context. Claude Tag goes a step further by treating the model as a persistent workspace presence rather than an on-demand lookup tool. Whether that changes how teams habitually reach for it is an open empirical question, but the architectural shift is real: the model now has a durable identity within the collaboration layer, not just a pipe into it.

The research preview label signals that Anthropic is gathering usage data before hardening the feature set. That framing is common enough in enterprise AI that it has become almost a standard shipping vehicle — but it also carries genuine meaning when a product is being placed inside asynchronous communication infrastructure. Slack threads can move fast, span large teams, and accumulate sensitive operational context. How Claude Tag handles permissions, what message history it can access, and how that access is scoped to specific channels or roles are the practical questions engineering and security teams will want answered before broad rollout.

Anthropic has not yet announced which other platforms will follow Slack in the Claude Tag rollout, though the announcement frames Slack as the starting point rather than the full scope of the feature.

The broader pattern — AI assistants embedded inside the collaboration tools where work actually happens, rather than siloed in standalone chat interfaces — has been playing out across the industry. Microsoft's Copilot sits inside Teams and the Office suite; Google's Gemini is woven through Workspace. Anthropic's approach with Claude Tag is broadly consistent with where the category is heading, though the company has historically positioned Claude's strengths around nuanced reasoning and extended context handling. Whether those characteristics express themselves differently inside a thread-based, asynchronous environment compared to a direct-prompt session is worth watching as the research preview generates real-world data.

For teams already running Claude in their workflows, the practical near-term question is adoption friction. A Slack-native presence lowers the context-switching cost substantially — no separate tab, no copy-pasting thread content into a chat window. That reduction in friction has historically mattered more than raw capability improvements in determining whether a tool gets used routinely or occasionally.

Claude Tag is available in research preview as of June 23, 2026, via Anthropic's announcement.