Wouldn’t you want to know how a dress or shirt will look on you before you buy it? (Yes, we do have stores for that.) But in the age of online shopping, Google is suggesting an AI-powered alternative.
In 2023, Google launched a feature that would let you see how an outfit looks on a series of pre-selected models. Now, Google is rolling out the same technique at its Google I/O developer conference… just on you.
It’s like an AI-powered fitting room to try on clothes. All you’ll need to do is have someone take a full-body photo (clothed, of course) against a neutral background, then upload it to Google. Google’s AI will then analyze the clothing, compare it to you, and voila! You’ve tried it on.
If you’re anything like my wife, that description sparked a twenty-minute monologue describing how women’s bodies vary and how women of an “identical” size vary in hips, bust, and so on. Google claims that this has been thoroughly considered; however, according to Vidya Srinivasan, who leads ads and commerce at Google.
“We spent a lot of time working with a custom image generation model that is particularly tailored for fashion, and it has a very deep understanding of 3D shapes,” Srinivasan told reporters in a Google I/O briefing. “It also has a very deep understanding of fabrics, just how fabrics will fold and stretch and drape on a wide range of bodies. So now, once you have an image of this, you probably want to try this on multiple different dresses. So you have a save feature right there, and then you can also share it with friends, because a lot of times you want another opinion about what actually looks good.”
Google is combining this new “try it on” feature with some of its existing search magic, like trying to find a rug using Google’s AI Mode that will work in a particular room. Google is also sending out its search agents, via agentic search, to find the cheapest price over time. These agents will do everything but buy it for you; you’ll need to click “buy” to make it yours.
Google is rolling out the virtual try-on technology as part of its Search Labs, beginning today.
