OpenAI launches ChatGPT shopping research to challenge Google’s growing AI shopping ecosystem

ChatGPT now has “Shopping Research”, a new capability that aims to do the heavy lifting when it comes to online shopping by providing users with personalized product recommendations.

ChatGPT’s Shopping Research feature will be available across its mobile, desktop, and web clients to all users on Free, Go, Plus, and Pro plans who are logged in. OpenAI says that for the holiday shopping season, “nearly unlimited” use of the feature is on the menu for users.

Shopping Research: Your Interactive Buying Guide

Shopping Research presents itself as an interactive, visual assistant when a user asks for shopping assistance in ChatGPT. The tool starts by prompting a user to detail what it is they want to buy. For example, this could be a lightweight but powerful laptop, or a gift for someone with specific requirements. ChatGPT can ask follow-up questions to get more context or details to better scope out the user’s intent.

It then goes on to perform a deep web crawl on reliable sites to fetch updated pricing, specs, and product reviews. Users can then browse this information and further interact with Shopping Research. For instance, once a list of products is returned, a user can like or eliminate an individual option from consideration, and mark it as “interested” or “not interested.” Depending on the user’s input, ChatGPT will then re-scope its search and generate a buyer’s guide that surfaces potential trade-offs, “informative” differences between various products, or “good value” options.

Focus on product details

OpenAI says Shopping Research can generate helpful recommendations particularly in product categories where purchase decisions involve more “nuance.” This covers areas like electronics, beauty, home & garden, kitchen appliances, and outdoor equipment. Essentially, this is an emphasis on technical and feature-heavy categories, where OpenAI will be matching or even surpassing its key competitors.

OpenAI says Shopping Research is a “mini” version of GPT-5, fine-tuned for shopping assistance

How does it work?

Shopping Research is powered by a lightweight, bespoke AI model. OpenAI says it is “trained to search trusted sites for relevant information and can compare data points across multiple sources.” It also says that Shopping Research was fine-tuned from GPT-5, making it a “mini” shopping-oriented version of the flagship model.

OpenAI has also indicated that Shopping Research conversations will not be shared with retailers or passed onto third-parties. Product results are also pulled from public data, rather than paid advertising sources, and OpenAI aims to cite verifiable sources whenever possible. However, as OpenAI admits, the tool can occasionally make mistakes or misreport pricing and availability. It suggests that users always double-check links and independently verify listings on retailer sites.

OpenAI has also said that it will be working on Shopping Research continuously over time. This can take the form of updating categories covered, refining the personalization engine, or possibly extending Shopping Research to allow users to checkout in real-time (this is already a possibility in areas where its “Instant Checkout” system is present).

Summary

Shopping Research is an add-on that empowers ChatGPT to do more than just answer questions or write text. By making product discovery a conversational, two-way, feedback-driven process, Shopping Research is set to make the holiday shopping season smarter, faster, and more personalized than ever.