Google's new AI shopping lets you try clothes on your own photo

Online shopping comes with a familiar headache: guessing whether a shirt will look ridiculous on you, and wondering if the price will crash twenty minutes after you hit purchase.

Google announced a massive shopping update inside AI Mode that merges its Gemini capabilities with the Shopping Graph. The highlight reel includes dynamic conversational research, agentic checkout that buys items on your behalf, and a virtual try-on tool that uses your actual photos.

Try on billions of clothes using your own photo

Google previously let shoppers see clothes on models of diverse body types, but it is now letting you be the mannequin. You can upload a single full-length photo of yourself to virtually try on tops, pants, skirts, and dresses.

Behind the tool is a custom fashion-specific image generation model. Google says the model understands how fabric interacts with the human body, specifically accounting for how materials stretch, fold, and drape across different poses. The feature is launching today in the U.S. as a Search Labs experiment, and you can save the resulting outfits or send them to friends for a gut check.

Multi-step search and an agent that buys for you

Browsing is also getting a structural overhaul. When searching in AI Mode, Google uses a mechanism called query fan-out. If you ask for a bag suitable for a trip to Portland in May, the system fires off multiple searches behind the scenes to deduce that you need weatherproofing and easy-access pockets, updating a side panel in real time.

Then comes the checkout agent. When browsing Google's catalog of over 50 billion listings, you can tap to track an item, specify your desired size and color, and name your price. Once a price drop triggers, confirming the alert lets you tap "buy for me." Google then navigates the merchant site, drops the item in the cart, and finalizes the transaction through Google Pay. Both agentic checkout and AI Mode browsing will roll out in the U.S. over the coming months.

Why it matters

Google is actively trying to transform its search engine from a referral directory into an end-to-end commerce engine. By letting users visualize garments on their own bodies and delegating the checkout slog to autonomous agents, it eliminates two of the biggest friction points that drive shoppers to Amazon.

We think the move is tactically smart, but it raises obvious operational hurdles. Virtual try-ons are still simulations that cannot account for fabric thickness, stiffness, or erratic brand sizing. On top of that, letting an AI agent buy items on third-party sites via Google Pay introduces thorny customer support questions if returns, shipping mix-ups, or unexpected restocking fees occur. Even so, it gives us a clear look at where agentic consumer web apps are heading next.