- Anthropic launched Claude Tag on June 23, 2026, letting teams assign tasks to an AI agent by tagging it directly in Slack channels, responses are visible to the whole team, not just the requester.
- Anthropic says its own product team generates around 65% of its code through an internal version of Claude Tag, though that figure comes from a self-reported, internally favourable case study.
- Enterprise controls arrived across the same week as the launch rather than after it, from Okta-managed connector access on June 18 to full Claude Desktop deployment on AWS, Google Cloud and Microsoft Foundry on June 22 and Trusted Devices on June 25, which will determine whether IT teams treat this as enterprise-ready or a pilot risk.
Anthropic is now building the case that its own product team writes around 65% of its code through an internal version of Claude Tag, the AI agent it launched in Slack on June 23, 2026. That is the most concrete signal the company has offered about the tool’s actual impact, and it comes from the most favourable possible source.
AI Joins the Team in Slack Channels
The design choice that makes Claude Tag different from a standard AI chatbot is visibility. Tag Claude in a shared Slack channel and the response is public: every team member sees what was asked and what the model produced. Claude also retains context across a conversation thread, so follow-up requests don’t require re-explaining the background. Anthropic describes this as “multiplayer AI,” according to the company, positioning Claude as a shared team resource rather than a personal assistant.
The launch builds on two existing Anthropic products. Claude Cowork, generally available since April 2026 after a January 2026 research preview, lets Claude access specified folders and autonomously read, edit and create files for multi-step tasks. Claude Tag is described by Anthropic as an evolution of Claude Code, the company’s coding-focused product, extending the model’s reach across teams rather than within a single developer’s session.
Automating Emails, Code and Data Workflows
The feature covers three practical areas: email handling, coding and data work. Recent enterprise deployments of Claude Desktop now include Microsoft 365 connectors, giving Claude access to mail and documents inside a company’s own environment. The 65% internal code-generation figure is the clearest illustration of scale Anthropic has offered, though it is a self-reported number from an obviously favourable case study. For data and documents, Claude Cowork handles the file access layer, reading and editing across specified folders without requiring a human to shuttle files back and forth.
Whether these results translate to other organisations is harder to verify from public material. Teams considering a rollout would be weighing Anthropic’s internal numbers against their own workflows, security requirements and the real cost of onboarding an AI into shared channels where mistakes are visible to the whole team. For a broader look at how AI assistants perform in collaborative settings, the findings on Google Gemini and Copilot’s productivity trade-offs are worth reading alongside this.
Enterprise Control and Scaled Productivity
Around the Claude Tag launch, Anthropic expanded deployment options for Claude Desktop to organisations using AWS, Google Cloud and Microsoft Foundry, adding centralised IT controls for data routing, identity management and policy enforcement. That expansion landed on June 22, the day before Claude Tag, part of a run of enterprise releases that also brought Okta-managed connector access on June 18 and Trusted Devices on June 25.
The practical question for most teams is sequencing: the productivity case for Claude Tag depends on how much you trust the guardrails around it. Anthropic shipped the enterprise controls across the same week as the feature rather than after it, a deliberate choice and one that will be tested by how Okta integration and Trusted Devices hold up in real deployments. If you’re thinking about where AI fits into your team’s existing tools, the Salesforce Einstein AI pivot is a useful reminder that enterprise AI rollouts rarely go in a straight line. Stay up to date with the latest AI developments at Auton AI News.
