Google is launching a new AI image generation app to Workspace that it’s calling Pics, and it has a new feature to try and reduce the hassle of iterating on AI images: Instead of having to write an entire prompt just to change one small aspect of an image, you’ll be able to click on what you want to change and leave a note about what you want to see, almost like leaving a comment in a Google Doc.

Pics is powered by a mix of Gemini and Google’s Nano Banana 2 image model. In a demo shown to reporters, a Google employee working on an invite for a child’s birthday party wanted to tweak individual parts of the invite. She clicked on an image of a cat, and then a little pop-up showed up where she left a note to change the cat image to one of a dog. She also clicked on the address shown on the card to make a direct tweak to what that text should say, and then asked Pics to make the changes. After a beat, the image was indeed updated — though in the specific demo example, the new “dog” still heavily resembled a cat.

Pics is rolling out initially as its own app on the web, first for what Google calls “trusted testers” and then later this summer for subscribers to Google’s AI Ultra plan. The plan, down the line, is to eventually incorporate Pics right into other Workspace apps so that users can make and edit images without leaving the app that they’re working from.