MICROSOFT JUST GAVE AI THE ABILITY TO OPERATE ANY COMPUTER SCREEN WITHOUT NEEDING A SINGLE API Microsoft published a Copilot Studio update on May 26 that should alarm every company still running legacy software it assumed was safe from automation. Computer-using agents are now generally available. These agents do not need an API or a formal integration point to interact with software. They see a screen the way a human does and act on what they find there. They click. They fill forms. They navigate menus. They pull data from any application with a visual interface. This matters because most enterprise software in use today was never built for programmatic access. Internal tools, decade-old custom platforms, systems that predate the API era entirely. Computer-using agents make all of it automatable without any changes to the underlying software. The same update ships real-time voice capabilities running below five hundred milliseconds of latency, a rebuilt workflow canvas with conditional branching and parallel execution paths, and deep integration with Microsoft’s Work IQ workforce analytics. The platform can now examine how employees spend their time and make decisions based on that data. This is not a chatbot feature. Microsoft is building agents that operate enterprise software the way a skilled person does, inside the compliance and security framework IT departments already trust. For the thousands of companies still running software from 2010, the automation wave just stopped asking for permission and started arriving. Keywords: Microsoft Copilot Studio, computer-using agents, enterprise AI automation, AI agents general availability