Anthropic Previews Claude Duet, a Proactive AI Agent for Professionals
Anthropic Previews Claude Duet, a Proactive AI Agent for Professionals
Anthropic has unveiled Claude Duet, a newly previewed AI agent app the company says is designed to work proactively across everyday professional tasks rather than simply responding to prompts. According to Anthropic, Duet is aimed primarily at professionals and small-business users, distinguishing it from more general-purpose personal assistants. Its debut places Anthropic in an increasingly crowded field of AI agent products, with early coverage from outlets such as TechCrunch and Thurrott framing the launch as part of a broader competitive race among major AI developers.
What Claude Duet Is Built On
Anthropic says Claude Duet is built on its Claude Opus 5.5 model. The company describes the app as capable of acting proactively on behalf of users, initiating steps across tasks rather than waiting for explicit instructions at each stage. These model and capability details come directly from Anthropic's announcement and support documentation, and as with most new product launches, they have not yet been independently tested or confirmed by outside parties.
Marketplace Integrations and Commerce Features
Anthropic's published materials state that Duet can connect to more than 6,000 connectors and plug-ins through the Claude Marketplace, enabling it to interact with a wide range of third-party tools and services. The company also says Duet can complete purchases using Stripe's Link or Shopify's Shop Pay, with user approval required before sensitive steps are carried out. A recurring theme in early coverage is that these figures and functional claims are vendor-asserted; the scale of the connector ecosystem and the reliability of commerce features have not been independently verified through third-party testing.
Safety and Privacy Claims
Anthropic has stated that Claude Duet will not trade securities or move money between accounts. The company also says it has implemented a new classifier intended to screen outgoing messages and block the sharing of sensitive information, such as credit card numbers or government identification numbers, by default. Many observers note that these are meaningful safeguards if they perform as described, but they remain company claims at this stage rather than conclusions backed by independent audits or regulatory review. A recurring consumer concern with proactive AI agents generally is the handling of sensitive financial and personal data, and Anthropic's stated protections should be understood as self-reported until outside verification becomes available.
Availability and Access
Claude Duet is currently offered as a research preview on iOS, Android, and the web through Claude.ai. Access is limited to Claude Pro and Max subscribers, and interested users must join a waitlist. This limited rollout means that independent testing and broader user feedback are still minimal, which in turn limits the ability of outside observers to confirm how the app performs relative to Anthropic's descriptions.
Competitive Landscape
Coverage of the launch has positioned Claude Duet against a range of other AI agent efforts, including offerings associated with Google's Gemini models, Meta's Muse, Microsoft's Autopilot, and OpenAI's Dots. These comparisons appear largely in early press summaries and are framed, in some cases, as a competitive race among major AI companies. It is worth noting that such competitive framing is largely derived from vendor statements and journalistic shorthand rather than from independent, side-by-side performance testing. As the agent category matures, a clearer picture of how these products actually compare is likely to depend on third-party evaluation rather than company announcements alone.