Anthropic has introduced Claude Duet, a personal AI agent designed to act proactively for human customers by taking autonomous actions to complete assigned tasks. The company describes the tool as a system that approaches users with items requiring attention, completes requested work, and continues operating even when the user steps away, while checking in before acting on their behalf.
Anthropic states that Claude Duet is built for professionals and small business owners who manage more responsibilities than they can easily track. The company attributes the agent's utility to its superior judgment, claiming it is smart enough to determine what to prioritize and which tasks require direct user input. To function effectively, the agent connects to a user's existing applications, including email accounts and calendars, to gather necessary context. Each morning, Claude Duet sends a daily brief outlining the day's most important items.
The agent carries over memory, skills, and connected apps from a user's existing Claude account, a feature Anthropic says minimizes setup work. Beyond daily briefs, Claude Duet handles administrative duties such as scheduling meetings and creating financial or sales forecasts. Anthropic also noted that the tool can build web apps, draft job postings, and make purchases on a business owner's behalf.
Security professionals have raised concerns about AI agents accessing sensitive information like emails and calendars, warning users to limit data sharing. In response, Anthropic highlighted built-in security measures intended to make the app safe to use. The company emphasized that the agent never directly sees a person's passwords or sign-in tokens.