ChatGPT Launches Virtual Try-On: AI Shopping Experience Gets a Major Upgrade

OpenAI is rolling out a set of new shopping features for ChatGPT, allowing users to virtually try on clothing and accessories using their own photos, and save favorite items to a "Favorites" list. This move signals a clear acceleration in ChatGPT's evolution — from a pure conversational tool to a fully integrated AI shopping assistant.

Virtual Try-On: Turning Your Photo Into a Fitting Room

At the heart of the new functionality is virtual try-on. After uploading a personal photo, ChatGPT can overlay clothing or accessories onto the user's image, generating a preview that closely approximates how an outfit would actually look on them. This directly addresses one of online shopping's most persistent pain points — the inability to judge whether clothes will fit or suit your style before placing an order.

Compared to the standard model photos provided by traditional e-commerce platforms, try-on previews based on the user's own image carry far more personalized value. For hesitant shoppers, this what-you-see-is-what-you-get experience has the potential to reduce return rates — and gives OpenAI a meaningful foothold in the high-value e-commerce space.

Virtual Try-On technology is not a new concept. Amazon, Snapchat, Pinterest, and others have all explored similar features, typically combining computer vision with generative AI — detecting body contours and pose keypoints, then fitting garment images onto the body according to perspective. Early implementations relied on AR (augmented reality) overlays, often producing stiff, unconvincing results. In recent years, the maturation of diffusion models has dramatically improved the naturalness of image blending — lighting, wrinkles, and occlusion can all be simulated through image generation rather than simple layer compositing. OpenAI's key advantage in integrating this capability into ChatGPT's conversational flow lies in GPT-4o's multimodal architecture, which can simultaneously process text and images. This enables a seamless end-to-end experience: a user says they're looking for a loose linen shirt → the AI recommends products → generates a preview of the user wearing it — all without jumping to a separate AR tool.

Favorites: From One-Off Conversations to an Ongoing Shopping Journey

Beyond virtual try-on, OpenAI has also introduced a "Favorites" feature. Users can save products they're interested in while browsing, building a personal item library for later comparison, review, or decision-making.

The significance of this design is that it moves ChatGPT's shopping experience beyond a single Q&A exchange and into a sustainable, cumulative usage loop. Users can gradually build out their shopping lists across multiple interactions — a concept borrowed directly from the "wish list" logic of established e-commerce platforms, but layered with the advantages of AI-driven conversational interaction.

The Strategic Play in AI-Powered Shopping

ChatGPT doubling down on shopping features reflects OpenAI's ongoing exploration of monetization pathways. Shopping recommendations are naturally monetizable — whether through referral revenue sharing, advertising, or transaction commissions, e-commerce is an unavoidable high-value traffic scenario for large model companies.

By combining computer vision capabilities like virtual try-on with conversational recommendations, OpenAI is effectively building a complete "AI personal shopper" experience: users describe what they need in natural language, the AI recommends products, generates try-on previews, and helps save and compare options — all within the chat window. This poses a genuine challenge to traditional e-commerce search and recommendation models.

Conversational Commerce refers to the process by which consumers complete the full journey — from product discovery to checkout — through natural language interfaces like chat or voice. The term was first coined by Chris Messina in 2015 and has since found practical application in platforms like WhatsApp Business and WeChat Mini Programs. Its core value lies in reducing the cognitive friction of shopping: traditional e-commerce depends on keyword search plus filters, requiring users to translate their needs into search terms; a conversational interface lets users describe their intent in vague, natural language and lets the AI handle intent parsing and product matching. Meanwhile, the rise of large language models means AI recommendations can now incorporate contextual memory and multi-turn clarification — theoretically better at capturing latent preferences than algorithmic recommendations. This is the underlying logic behind Google, Perplexity, and other search and AI companies all staking out positions in the shopping space — whoever becomes the user's first point of entry for purchase decisions controls an enormously valuable traffic node.

Questions Worth Watching

Despite the promising outlook, several key questions deserve attention. First is the accuracy of virtual try-on — whether AI-generated outfit previews can faithfully convey fit, fabric, and drape will directly determine user trust. Second is the privacy question: uploading personal photos involves sensitive data, and how OpenAI handles and protects these images will be a critical consideration. Third is the completeness of the product sourcing and transaction loop — for now, the feature skews more toward the browsing and preview stage than a full end-to-end commerce experience.

Taken together, the addition of virtual try-on and Favorites meaningfully strengthens ChatGPT's practical utility in consumer scenarios, and offers a compelling case study in how conversational AI and e-commerce can be brought together.