NEW YORK, April 17: Microsoft has rolled out a new “computer use” feature in Copilot Studio this week, allowing AI agents to interact directly with websites and desktop applications. Similar to OpenAI’s Operator and Claude’s feature of the same name, this functionality enables businesses to build AI agents that use websites and apps just like a human would—automating tasks without requiring an API.

“Computer use enables agents to interact with websites and desktop apps by clicking buttons, selecting menus, and typing into fields on the screen,” explained Charles Lamanna, Corporate Vice President of Microsoft’s Business & Industry Copilot. “This allows agents to handle tasks even when there is no API available to connect to the system directly. If a person can use the app, the agent can too.”

With this new capability, AI agents created through Copilot Studio will be able to automate tasks such as data entry, market research, and invoice processing. Microsoft notes that these agents will also adapt to changes in app or website layouts—such as moving buttons or altered screens—allowing them to continue functioning without interruption.

Earlier this month, Microsoft introduced a similar feature called "Actions" to its consumer version of Copilot. That feature allows background automation, enabling users to book restaurant reservations, purchase event tickets, or shop online without interrupting their workflow. While Actions is currently limited to select partners, Copilot Studio’s "computer use" functionality is expected to work across a broader range of websites and applications.