Copilot has learned to see your screen: a new feature in Edge

Microsoft is taking another step toward integrating artificial intelligence into everyday browser use: the Copilot Vision feature is now available to all Edge users – for free and without a subscription. With this option, Copilot is able to literally analyze the content of an open website and help in real time. The launch of the new feature was announced by Microsoft AI CEO Mustafa Suleiman on the Bluesky platform.

Mustafa Suleiman, CEO of Microsoft AI, announced the launch of the new feature on the Bluesky platform.

Copilot Vision is an interactive voice assistant that can “see” everything on the screen and tell you what to do next. It’s like a digital satellite that can not only understand context, but also adapt to the user’s tasks.

Copilot Vision is an interactive voice assistant that can see everything on the screen and tell you what to do next.

The feature only works after voluntary connection: the user must manually agree to activate Copilot Vision. After that, the assistant can accompany you while you cook – guiding you through the steps of a recipe, or, for example, helping you parse a job description and even formulate interview responses or draft cover letters. Though Microsoft ironically notes that using AI to write resumes is a controversial idea.

Officially, Vision can highlight parts of a page to help you navigate faster, but it doesn’t perform actions for the user – it doesn’t open links or click buttons. It’s purely an observation and guidance system.

The deeper integration of Vision into the system remains the domain of the paid Copilot Pro subscription for now. With the Pro version, the AI can interact not only with the browser, but also with desktop applications, from Photoshop to video editors and even games. For example, Vision recently demonstrated how it prompts a Minecraft walkthrough in real time.

An AI can also interact with desktop applications, from Photoshop to video editors to even games.

How to enable Copilot Vision in Edge

To try out the new feature, you’ll need to open Microsoft’s website in the Edge browser, where you’ll be prompted to activate Vision. Once you agree, a microphone icon will appear in the Copilot sidebar – that’s where the interaction session begins. When you start, the browser changes its color scheme and a sound signals the start of the interaction.

But, as the reporters note, the process can be choppy: sometimes the browser takes several attempts to offer activation, and in some cases the control panel never appears. On older devices with limited resources, the launch may be error-prone.

But the process can be unstable, with the browser sometimes taking several tries to offer activation and sometimes the control panel never appearing.

Microsoft emphasizes that while Vision is running, the company only captures Copilot responses and does not collect information about page content, images, or user actions. You can disable the feature at any time by simply ending the session or closing the browser window.