OpenAI is introducing a new shopping research experience inside ChatGPT that turns product hunting into a guided conversation instead of a long tab-by-tab search. The feature is starting to roll out on web and mobile for logged-in Free, Go, Plus, and Pro users, with nearly unlimited usage through the holidays to support gift buying and big-ticket decisions.
A New Shopping Copilot In ChatGPT
With shopping research, users can describe what they need in natural language, from a “gaming laptop under $1000 with a screen over 15 inches” to a “full-length dress under $250”. ChatGPT then asks follow-up questions about budget, who the product is for, and which features matter most before starting its research.
Behind the scenes, it looks across the internet for up-to-date information on prices, availability, specs, reviews, and images, then returns a curated set of options with clear differences and tradeoffs instead of a single quick answer.
The new mode is built for deeper, decision-heavy journeys, while regular ChatGPT replies still handle simple checks like confirming a feature or price. OpenAI notes that hundreds of millions of people already use ChatGPT to understand and compare products, and this experience is tailored to that use case.
Guided Discovery And Lookalikes
Shopping research shines in categories where details matter, including electronics, beauty, home and garden, kitchen and appliances, and sports and outdoor gear. It can, for example, help someone compare several stroller models based on weight, foldability, comfort, and terrain, or track down a dress that looks like a reference image but fits a different length or price point.
As suggestions appear, users can mark items as “Not interested” or “More like this,” helping the system learn what to prioritise and what to filter out in real time.
The experience also supports deal hunting, such as finding the best Black Friday price on a specific sneaker and checking whether a student discount might apply. When shoppers are ready to buy, they can click through directly to the retailer’s website, with Instant Checkout planned for the future so purchases can be completed inside ChatGPT for participating merchants.
How Shoppers Use The New Interface
To start, users either ask a shopping-focused question or select the shopping research option from the ChatGPT menu. A dedicated visual interface opens, where chat messages sit alongside product cards, giving a more familiar shopping feel while keeping the back-and-forth flow.
Shoppers answer short questions about budget, style, intended user, and priorities, and can refine as they go, rather than filling out a rigid form or starting new searches for every tweak.
After a few minutes, shopping research delivers a personalised buyer’s guide that summarises top products, key differences, and tradeoffs, with context that would normally require reading multiple product pages and reviews. The guide is designed to cut down the time it takes to move from “just browsing” to a short, confident shortlist.
Powered By GPT‑5 Mini For Shopping
The new experience runs on a version of GPT‑5 mini trained with reinforcement learning specifically for shopping tasks, so it can better handle structured product data and detailed specs. OpenAI trained this model to read trusted sites, pull accurate product information, cite sources correctly, and synthesise what it finds into detailed, concise research.
The model is also designed to refine its answers as shoppers add or change constraints, such as adjusting the budget or preferring a different brand, without forcing them to restart the process.
Transparency, Trust, And Limitations
Shopping research is built to be transparent: chats are not shared with retailers, and results come from organic, publicly available retail sites rather than sponsored placements. Merchants that want to appear can follow an allowlisting process so the system knows it can safely crawl and reference their sites.
Even with this shopping-tuned model, OpenAI notes that there can still be errors in details like price and availability, and encourages shoppers to verify final information on the retailer’s page before checking out.
