Claude Duet turns chat into ongoing collaboration
Claude Duet is Anthropic’s proactive AI agent rather than a chatbot that simply waits for prompts. It can prepare a morning brief from connected work signals, identify items that need attention, and accept tasks by text or voice. It then continues working on a cloud computer and returns when the work is finished or a decision is required.
What Claude Duet can handle
The examples include scheduling meetings, building dashboards, preparing fundraising materials, drafting emails and contracts, and producing a business pulse from point-of-sale, Shopify, CRM, support, payroll, and banking data. Anthropic says it works with more than 6,000 connectors and plugins and carries over the user’s Claude memory, skills, and apps. Sensitive actions such as sending messages or making payments still require approval.
The broader shift is from one-off answers to continuing task orchestration: users set an objective, then review outputs or important checkpoints. Anthropic says approvals, logins, and safety checks run in the user’s own cloud computer, while activity remains visible for steering. Those controls, however, do not make the results inherently reliable.
Claude Duet’s limits and availability
Claude Duet is currently a research preview on a waitlist for Claude Pro and Max subscribers in the United States; Team and Enterprise plans are not included. Background checks share the account’s normal usage limits, and larger tasks consume more model turns. Anthropic also warns that the agent may misread a thread, misjudge urgency, or miss something that arrives between check-ins. The key open question is therefore not whether it can complete polished demos, but whether people can supervise it effectively across complex, ongoing, and higher-stakes work.