ChatGPT Introduces Shopping Research

OpenAI has introduced a new feature in ChatGPT called shopping research, designed to create personalised buyer’s guides by gathering product information from across the web. This tool is now accessible to all logged-in users on mobile and desktop, including Free, Go, Plus, and Pro plan subscribers. OpenAI has made the feature available with almost unlimited usage throughout the holiday season.

How Shopping Research Works

Unlike standard ChatGPT responses, shopping research interacts with users in a more guided way. Users describe what they are looking for and answer a few clarifying questions regarding their preferences and budget. Within a few minutes, the system generates a tailored buyer’s guide.

The feature compiles a range of product details, including prices, availability, reviews, specifications, and images, all sourced from the web. Users can influence the results by marking items as “Not interested” or “More like this,” helping the tool refine its recommendations in real time.

According to OpenAI, shopping research is intended to make product discovery more conversational and decision-focused. It asks the right questions, collects up-to-date information from high-quality sources, and presents results that users can customise according to their needs.

Best Categories for Shopping Research

The tool is reported to perform especially well in categories such as electronics, beauty, home and garden, kitchen appliances, and sports and outdoor equipment. OpenAI highlights that these categories benefit the most from the system’s data aggregation and personalisation capabilities.

Technical Details

Shopping research is powered by a GPT-5 mini variant specialised in shopping tasks, which is post-trained on GPT-5-Thinking-mini. Internal testing by OpenAI revealed that the tool achieved 52% product accuracy on complex queries with multiple constraints, compared with 37% for ChatGPT Search.

Product accuracy measures how effectively responses meet user requirements, including price, colour, material, and specifications. OpenAI designed the system to update results in real time, learning from user feedback to improve accuracy and relevance.

Privacy and Data Sharing

OpenAI emphasises that user interactions with shopping research are private and never shared with retailers. The results are generated organically using publicly available retail websites. Merchants interested in appearing in the shopping research results can participate through an allowlisting process provided by OpenAI.

Limitations

While the feature offers a streamlined product discovery experience, OpenAI acknowledges it is not without imperfections. The system may occasionally provide inaccurate product details, such as pricing or stock availability. Users are encouraged to verify critical information on retailer websites to ensure accuracy before making purchasing decisions.

Why This Matters

Shopping research brings more of the early product comparison journey directly into ChatGPT. Users can now explore options and make informed choices without leaving the platform for traditional search engines.

For retailers and affiliate marketers, this development raises the stakes. Product visibility in ChatGPT may depend on how well their items and webpages are represented in OpenAI’s system, including participation in the allowlisting process.

Looking Ahead

The feature is now live for all logged-in users. OpenAI has indicated plans to introduce direct purchasing through ChatGPT for merchants using Instant Checkout, though no exact timeline has been provided. This could further streamline online shopping by reducing the need to navigate separate websites for purchases.