SAN FRANCISCO: Anthropic has officially launched Claude Cowork for all paid users, moving it out of its testing phase. Announced on April 9, the update introduces six new enterprise-focused features aimed at helping companies use the tool at scale.
Claude Cowork is an AI assistant designed for non-technical users. Available on macOS and Windows through the Claude desktop app, it can handle tasks like creating documents, updating projects, doing research, and preparing collaboration materials, allowing users to focus on other work. Early adoption shows strong usage across operations, marketing, finance, and legal teams.
To make the tool enterprise-ready, Anthropic has added new controls for IT teams. These include role-based access, where admins can group users and decide what features each group can use. Integration with identity systems (via SCIM) makes it easier to manage access across organizations. Companies can also set spending limits for different teams or departments.
Another key update is more detailed control over integrations. For example, with Gmail, admins can allow reading emails but block sending them. This level of control addresses a major concern for enterprises around security and compliance, which had previously slowed adoption.
Anthropic also introduced a Zoom integration that brings meeting summaries, transcripts, and action points directly into workflows. Additional features like usage tracking and OpenTelemetry support provide better visibility and monitoring.
Early users of Claude Cowork include companies like Zapier, Jamf, and Airtree.
In the competitive landscape, Claude Cowork goes up against tools like Microsoft Copilot, Google Gemini (Agent Mode), and OpenAI Operator.
Alongside this release, Anthropic also introduced Claude Managed Agents in public beta, APIs that allow developers to build and run AI agents in the cloud.
Previous
