OpenAI began rolling out shopping features for ChatGPT on October 1, 2026, and the headline capability is one that online apparel retailers have chased for years: the ability to see a garment on your own body before you buy it. According to TechCrunch, the new tools let users virtually try on clothing and accessories using their own photos, then store the products they like in a Favorites library inside the app. The feature arrives not as a standalone fashion app but as an extension of the assistant millions of people already use for everything from résumé drafting to recipe planning.
ChatGPT's New Virtual Try-On Feature Explained
Fit-related returns cost U.S. retailers an estimated $100 billion or more each year, according to widely cited industry analyses of e-commerce return data, and apparel consistently ranks as the most-returned product category online. That single statistic explains why a virtual try-on tool inside ChatGPT matters far beyond novelty. The mechanic, as TechCrunch reports it, is straightforward: a user supplies their own photo, and ChatGPT renders how clothing or accessories would look on them. It is a personalization layer built on top of the assistant's existing conversational interface, rather than a separate shopping destination.
The practical appeal is obvious to anyone who has ordered two sizes of the same shirt knowing one will go back. Traditional product photography shows a garment on a model whose proportions rarely match the shopper's. Virtual try-on attempts to close that gap by substituting the shopper's own image for the model's. Whether the rendering is accurate enough to trust is the open question — and the one that will determine whether the feature changes buying behavior or simply becomes a fun detour.
Users do not need to learn a new app or create a separate account. The try-on experience lives where conversations already happen, which lowers the friction that has historically slowed adoption of dedicated AR shopping tools.
The ChatGPT Favorites Library for Shopping
Saving a product in most shopping apps means adding it to a cart you may never check out, or bookmarking a page you will forget. OpenAI's approach is a Favorites library — a place inside ChatGPT where users collect products they are considering, per TechCrunch's report. The naming is deliberate. A cart implies imminent purchase; a favorites list implies ongoing consideration. That distinction matters for a tool meant to sit alongside browsing and research rather than replace the checkout flow entirely.
Read next Laika's Wildwood: Stop-Motion Fantasy at TIFF 2026The library also solves a problem specific to conversational AI. Without persistent storage, every product discussion would vanish when the chat window closed, forcing users to start over. A Favorites library gives the assistant memory of what a shopper liked, which in principle allows for follow-up comparisons, outfit suggestions, or price checks in later sessions. OpenAI has not detailed how those follow-ups will work, and the summary of the rollout does not specify which retailers or brands are participating. Those gaps are worth watching, because a try-on tool is only as useful as the catalog behind it.
Why OpenAI Is Entering the Shopping Space
Shopping queries are among the most commercially valuable searches on the internet, and they represent one of the few categories where Google has historically been almost impossible to displace. When OpenAI pushes ChatGPT toward product discovery and virtual try-on, it is chasing advertising and commerce revenue that currently flows overwhelmingly to search engines and marketplaces. The company has spent years building consumer habit around its assistant; converting that habit into purchase intent is the logical next step.
There is also a defensive logic. If users begin asking ChatGPT what to buy — and many already do — OpenAI captures the top of the purchase funnel whether or not it ever handles the transaction. The try-on feature and Favorites library are ways to deepen that role. Once a shopper has uploaded a photo and saved a dozen items, switching to a rival assistant carries a real cost. That is the same lock-in dynamic that made saved payment methods and wish lists so sticky for Amazon and Google.
How AI Virtual Try-On Compares to Existing Solutions
ChatGPT's entry does not arrive in a vacuum. Google has integrated virtual try-on for apparel into its Shopping results, letting users see clothes on a range of body types. Amazon has experimented with similar tools, including virtual shoe try-on and outfit visualization, tied directly to its retail catalog. Dedicated players such as Zeekit — acquired by Walmart — and a cluster of fashion-tech startups have offered AR fitting rooms for years. Grand View Research has projected the global virtual try-on market to grow at a double-digit compound annual rate through the late 2020s, reflecting steady investment across retail.
What ChatGPT adds is distribution and conversational context. Google's try-on lives inside a search results page; Amazon's lives inside a store. ChatGPT's version lives inside a dialogue, where a user can ask for a blazer, refine the request by occasion or budget, and see options on their own image without leaving the thread. That conversational wrapper is genuinely different from a grid of product tiles.
What ChatGPT lacks, at least initially, is the thing incumbents have spent years building: deep inventory integrations, accurate sizing data, and return infrastructure. Fashion-tech analysts have long argued that try-on accuracy depends less on the rendering model than on the quality of garment measurements and fit data feeding it. A persuasive image that misrepresents how a jacket drapes will increase returns, not reduce them — the opposite of the feature's stated promise.
Implications for the Future of AI-Powered Fashion Shopping
If virtual try-on works, the downstream effects touch nearly every part of apparel retail. McKinsey research on returns has repeatedly found that sizing and fit issues drive the majority of apparel returns in online channels, with rates in some categories exceeding 30 percent. Even a modest reduction in that figure translates into meaningful savings in reverse logistics, restocking, and markdowns — costs that ultimately show up in the prices shoppers pay.
There is a consumer-side benefit too. Return shipping is not free in any real sense; it costs time, packaging, and often a fee. A shopper who buys with more confidence because they saw the item on their own body is a shopper who keeps more of what they order and returns less.
The risks are equally concrete. Virtual try-on depends on users handing over personal photographs and, potentially, body measurements. How OpenAI stores, uses, and deletes that imagery — and whether it becomes training data — will be a central privacy question. The reported rollout summary does not address data handling, which means scrutiny is likely to arrive before clarity does. There is also the hallucination problem: an AI that renders a garment slightly wrong is not just unhelpful, it is actively misleading in a way that costs the shopper money.
What This Means for Everyday Shoppers
For now, the practical takeaway is modest. If you already use ChatGPT, a try-on option will appear in your shopping conversations, and you can save items you are considering to a Favorites list. Treat it as a second opinion, not a guarantee. Virtual renderings cannot replicate fabric weight, stretch, or how a collar sits after a wash — the reasons most people still want to touch a garment before committing.
The feature is a signal about where AI assistants are heading. The same conversational layer that helps you plan a trip or debug a spreadsheet is being pointed at your wardrobe, with your photo as the input. That is a meaningful shift in how purchase decisions get made, and it will play out not in a single launch but over the next several shopping seasons. Whether ChatGPT virtual try-on becomes a routine step in how people buy clothes depends on one thing: whether the images it produces are trustworthy enough that shoppers stop ordering two sizes just to be safe.
Source: TechCrunch
