Anthropic launched Claude Duet, an AI agent powered by Claude Opus 5.5 that manages workplace tasks across connected apps. It handles scheduling and dashboards while keeping data safe with a new security system. Anthropic said, "Claude Duet is still an early version," and opened a waitlist for eligible users today.
Powered by Claude Opus 5.5, the new AI agent can send daily summaries of important events across connected apps.
It also monitors these apps throughout the day and alerts users only when something requires their attention.
The tool can handle a wide range of tasks, from simple ones like scheduling meetings to more complex ones like creating dashboards with multiple tools or financial forecasts.
Claude Duet also comes with a virtual machine and browser for completing tasks. It draws on memory, skills, and connected apps from Claude.
Before taking any major action, such as making a payment or deleting something, the AI agent asks for user approval.
To protect sensitive information, it also has a new "dedicated classifier and safety system prompt" that prevents requests to share sensitive data like credit card numbers.
Anthropic has clarified that Claude Duet is still an "early version" and could misinterpret a thread or misjudge what's urgent.
The company has opened a waitlist for the AI agent, which is available to eligible users during this preview phase.
However, personal tasks such as booking tickets aren't its strong suit yet, as it's primarily designed for workplace use.