ChatGPT gets a personal shopping researcher

Finding the right product usually means opening thirty browser tabs, dodging sponsored reviews, and guessing which specs actually matter. OpenAI wants ChatGPT to do that legwork for you.

OpenAI introduced shopping research in ChatGPT, a dedicated tool designed to turn product discovery into an interactive, step-by-step buyer's guide.

The feature is rolling out on web and mobile to all logged-in Free, Go, Plus, and Pro users. To catch the holiday rush, OpenAI is offering near-unlimited access across all tiers through the holidays.

An interactive shopping assistant

Instead of just spitting out a static top-five list, shopping research opens a visual interface that treats shopping like an ongoing conversation. You can trigger it from the (+) menu or by asking a comparison question, such as picking between three gravel bikes or finding a quiet cordless vacuum.

The tool kicks off by asking clarifying questions about your budget, intended use, and must-have features. If you have ChatGPT memory enabled, it factors in past context—like knowing you play PC games when suggesting a new laptop.

While you chat, the system browses the web for real-time prices, availability, specs, and user reviews. As options pop up, you can tap buttons like "More like this" or "Not interested" to steer the search. Within a few minutes, it synthesizes everything into a custom guide detailing key differences and tradeoffs.

For now, clicking an item sends you directly to the retailer's site. OpenAI noted that in the future, users will be able to complete purchases directly inside ChatGPT with select merchants through an upcoming Instant Checkout feature.

For ChatGPT Pro subscribers, the capability also hooks into Pulse, meaning the AI might proactively serve up personalized shopping cards based on previous topics—like accessory recommendations days after you discussed e-bikes.

RL-tuned GPT-5 mini behind the counter

Under the hood, shopping research runs on a custom version of GPT-5 mini, fine-tuned with reinforcement learning specifically for commerce and sourcing tasks. OpenAI says it was trained to verify specs, cite reliable retailers, and ignore low-quality or spammy review sites.

The company emphasized that user chats are never shared with merchants and that recommendations remain strictly organic rather than paid placements. Retailers can join an allowlist to ensure their listings appear in results.

OpenAI acknowledged that the model can still occasionally misquote a flash sale price or inventory level, advising shoppers to verify final details on the merchant's checkout page.

Why it matters

We think this is OpenAI's most direct strike against Google's core cash cow to date.

When buying high-consideration items—like appliances, beauty products, or outdoor gear—users spend hours sifting through affiliate blogs and ad-cluttered search results. By condensing product discovery, spec comparison, and soon checkout into a single interface, ChatGPT removes nearly all that friction. For shoppers, it is well worth testing for upcoming big-ticket holiday purchases.

The strategy here is sharp: hook users with an ad-free, objective research experience right before the biggest spending season of the year, while quietly laying the plumbing for Instant Checkout. If OpenAI becomes the starting point where people decide what to buy and where they pay, it transforms ChatGPT from a productivity helper into an end-to-end commerce powerhouse.

That puts immense pressure on Google and Amazon. Google lives on commerce search ads, while Amazon relies on shoppers starting their search within its own app. If conversational agents prove they can curate better than an algorithm loaded with sponsored listings, the traditional search-and-buy pipeline is going to look outdated very fast.