OpenAI is making another run at turning ChatGPT into a shopping assistant. On Thursday, the company announced the global launch of two shopping features — a virtual try-on tool and a product favoriting system — as it works to keep users inside its chatbot from first search to final purchase, according to TechCrunch.
The rollout is OpenAI's most substantial shopping push since an earlier instant-checkout experiment stalled, and it lands in a market where Google has offered a virtual try-on of its own for about a year. It is also the latest sign that conversational AI assistants are quietly becoming retail storefronts in their own right. For more on stories like this, see our breaking AI news coverage.
How Virtual Try-On Works
The new feature lets ChatGPT users upload a selfie or a full-body photo to visualize how an article of clothing or an accessory might look on them. The option appears as a Try On button in ChatGPT's shopping results, according to the company.
Users can also work in the opposite direction: upload an image of an item — a web screenshot, for example — and ask ChatGPT to try it on for them. That turns the chatbot into a fitting mirror for products discovered anywhere on the web, not just items surfaced by ChatGPT's own shopping results.
Alongside try-on, OpenAI is launching Favorites, a way to save products users discover to a Library in the app so they can return to them later. The company notes that saved items will be stored alongside users' try-on images, keeping the whole consideration process in one place.
Powered by ChatGPT Images 2.5
Both features lean on ChatGPT Images 2.5, the image generation model OpenAI launched in September. The company claims the new model produces more natural lighting and richer textures, follows editing instructions more reliably, and reduces image generation latency — all qualities that matter when the output is a simulated photo of clothing on a customer's own body.
The model upgrade is central to whether the feature convinces shoppers. A virtual try-on that renders fabric inaccurately or distorts fit does not just fail aesthetically; it undermines purchase confidence. OpenAI's claims about improved realism have not been independently verified, but the company is clearly betting that image quality has crossed the threshold where simulated fitting is useful rather than gimmicky.
OpenAI's Second Attempt at Shopping
The launch is also a course correction. As TechCrunch notes, OpenAI previously experimented with an instant checkout feature that ended up not performing well, and the company has had to pivot away from it.
The shopping-assistant space has proven tricky for AI companies more broadly. Agentic AI startup Instinct recently began pushing proactive product recommendations to its users, and some recipients felt the suggestions were overreach — more akin to ads than helpful assistance. The episode illustrates the delicate line consumer AI companies must walk: users appear willing to let agents shop when asked, but unsolicited commercial suggestions quickly read as advertising.
OpenAI's try-on and Favorites are deliberately user-initiated rather than proactive. The features respond to explicit shopper intent — visualize this on me, save this for later — rather than pushing products unprompted, which may help the company avoid Instinct's reception.
Competing With Google and Pinterest
The competitive context matters. Google launched virtual try-on last year, giving it a twelve-month head start in refining the experience and in persuading retailers to participate. Pinterest and Google have dominated the fashion inspiration and discovery space in recent years, building businesses on the moment when a saved look converts into a sale.
ChatGPT's use cases aim straight at that behavior. OpenAI says users could describe a style they want to try and ask the assistant to shop for the pieces needed to complete the look. Or they could upload photos of celebrities' outfits and ask ChatGPT to find the items those celebrities are wearing that are available for purchase. Both flows replicate — inside a chat window — the screenshot-and-search ritual that has made Pinterest and Google Lens fixtures of online fashion discovery.
The Open Question
Whether ChatGPT becomes a first stop for fashion discovery, rather than a novelty, remains to be seen. Google's year of experience with try-on suggests the feature is easier to launch than to make habitual, and shopping habits are among the stickiest consumer behaviors on the internet.
Still, the strategic logic for OpenAI is clear. Shopping queries bring users back daily, generate commercially valuable intent, and give retailers a reason to care about how their products appear inside the assistant. With Favorites, OpenAI now captures the save-for-later moment; with try-on, it captures the fitting room. What it does not yet capture — after the instant-checkout retreat — is the transaction itself. For now, OpenAI appears content to own discovery and deliberation, the most valuable steps short of the buy button.
---
Stay Ahead of AIGet the latest AI news, analysis, and breakthroughs — all in one place.
Read more AI news →
