Anthropic launches Claude Tag for Slack, enabling AI task delegation

Anthropic has unveiled Claude Tag, a new Slack‑native feature that turns the company’s language model into a persistent, team‑level assistant. The beta release is available to users on the Claude Enterprise and Team plans and allows team members to tag @Claude in any channel, assigning it tasks that the model can remember, schedule, and complete over time. By treating Claude as a teammate rather than a one‑off chat partner, the feature supports asynchronous updates, enabling users to receive progress reports without interrupting their workflow.

The core of Claude Tag is its memory and task‑persistence engine. Once a task is assigned, the model retains the context of the channel, the relevant data sources, and any code repositories it has access to, eliminating the need to re‑explain the objective in subsequent interactions. The system also integrates with Slack’s existing tool ecosystem, allowing administrators to grant selective permissions and set spend limits on the model’s use of external APIs. This design gives teams fine‑grained control over cost, security, and compliance while still benefiting from AI‑driven automation.

Internal usage data underscores the feature’s impact: the Claude Code team reports that Claude Tag now writes roughly 65 % of the product team’s code, including the majority of the work that built the feature itself. The beta status encourages controlled testing, but early adopters are already using the tool to automate routine tasks such as chasing product metrics, managing support tickets, and debugging code, freeing human resources for higher‑value work.

Anthropic plans to extend Claude Tag beyond Slack to other collaboration platforms, but the current focus remains on Slack because of its ubiquity in team environments. The launch signals a shift toward embedding AI agents directly into workplace tools, positioning them as assistants rather than replacements.