What Happened
On October 1, 2026, ChatGPT Try On began rolling out globally in ChatGPT shopping results. Users can upload a selfie, a full-body photo, or a product screenshot, and the Images 2.5 model generates the outfit on them. Results are saved to a Library for later reference.
It is part of a wider shift toward AI-led shopping. The same day, Albertsons was reported to be using ChatGPT Enterprise internally and the OpenAI API for customer experiences across more than 2,200 stores, with a Safeway chatbot in ChatGPT turning recipes, photos, or requests into product suggestions, carts, and in-store navigation.
Three Takeaways for Brands
- Product imagery is now input for AI. Clean, well-lit, consistent product photos on every angle give AI shopping tools better material to work with. Messy images risk looking worse in a generated try-on.
- Discovery is moving into the chat. Shoppers may compare, visualize, and shortlist inside an AI assistant before they ever open your site, so your product data must be accurate and easy to read.
- Trust signals still decide the purchase. A generated image helps people picture the item, but sizing details, reviews, and clear policies are what turn that into a checkout.
Try On lowers the imagination gap for fashion buyers. It also raises the bar: if your catalog data is incomplete or inconsistent, an AI shopping assistant has less to show and less reason to recommend you.
What You Can Do This Week
Audit your top ten products: check photo quality, sizing information, and descriptions for accuracy and consistency across your website and marketplaces. Then fix the gaps so your catalog is ready for AI-assisted shopping.
