Anthropic introduced Claude Tag, a new tool based on the Opus 4.8 model that integrates the network into Slack channels as a full team member. Users can set tasks via @Claude, after which the AI will split the task into stages and execute it in the background with access to databases, code, and third-party applications. The service is already available in beta for Enterprise and Team plans, replacing the standard Claude app for Slack.

The key features of Claude Tag are multi-user mode, context accumulation of discussions and proactivity: AI can itself remind about tasks and report updates. Moreover, the agent has a high degree of autonomy, allowing long-term projects to be planned and executed over several days. Within Anthropic itself, the tool already generates 65% of the software code and is also used extensively to analyze metrics and process IT requests.

To ensure security, administrators have the ability to restrict the network’s access to sensitive data and set limits on token consumption. The launch of the corporate AI agent from Anthropic comes against a backdrop of growing interest in integrating neuresets into business processes: earlier, in early June, its own virtual assistant for providing investment advice introduced the cryptobureau Coinbase.