Microsoft just gave Copilot a serious upgrade. It’s called Copilot Vision, and it’s starting to roll out in Windows Insider builds.

Copilot can now actually see what’s on your screen—images, PDFs, apps, all of it.

Instead of just answering questions you type in, it can respond to what you're already working on.

Ask it about a photo, get help with a doc that’s open, or pull info from something on your desktop—no copy-paste needed.

Copilot Vision runs on AI models that can understand visual context—so it’s not just reading text, it’s processing what’s on your screen like a human would.

It blends that with Microsoft’s existing tools to connect what you see with what you’re asking, making the assistant way more intuitive.

In a quick demo, it spotted ingredients in a photo and gave useful follow-ups.

It also understood content in a PDF and made connections across the screen. Definitely a step up from the old version.

Right now, it’s limited to a small group of testers, but Microsoft says more people will get it soon.

If it works as promised, it could actually make Copilot a solid everyday tool instead of something you mostly ignore.

Keep an eye out—especially if you’re already in the Windows Insider program. This one might be worth trying.