ChatGPT incorporates shopping assistant with personalized guides

OpenAI has introduced Shopping Research, a new feature integrated into ChatGPT that generates personalized shopping guides to help users with complex purchasing decisions. The tool began rolling out on November 24 for registered users on Free, Go, Plus, and Pro plans, both on web and mobile versions. During the holiday season, the company offers nearly unlimited use of this feature across all plans.

Unlike traditional ChatGPT responses, Shopping Research opens a visual interface where users describe what they're looking for through instructions like "find the quietest cordless stick vacuum for a small apartment" or "I need a gift for my four-year-old niece who loves art." The system asks clarifying questions about budget, recipient, or priority features before starting the research.

The feature uses information from previous conversations and ChatGPT memory to personalize results. During the search process, which takes a few minutes, the system collects updated information on prices, availability, reviews, specifications, and product images. Users can guide the research in real time by marking products as "not interested" or "more like this," allowing the system to adjust results based on expressed preferences.

The final result is a shopping guide with top products, key differences, pros and cons of each option, and updated information from reliable retailers. The tool performs especially well in categories with many technical details such as electronics, beauty, home and garden, kitchen and appliances, and sports and outdoor activities.

Shopping Research is powered by a version of GPT-5 mini trained specifically for shopping tasks through reinforcement learning. OpenAI trained the model to read trusted sites, cite reliable sources, and synthesize information from multiple sources to produce quality product research. The system is designed to update and refine its research in real time, incorporating new constraints and adjusting to user preferences.

Users can access Shopping Research when ChatGPT automatically suggests it after asking a shopping-related question, or by manually selecting the option from the menu. For ChatGPT Pro users, the feature is also available in ChatGPT Pulse, which can proactively suggest shopping guides based on previous conversations.

OpenAI notes that results are organic and based on publicly available retail sites, reading product pages directly and avoiding low-quality sites. User conversations are never shared with retailers. If users wish to purchase a product, they can access the retailer's site through links. In the future, OpenAI plans to enable purchases directly through ChatGPT for merchants participating in Instant Checkout. The company warns that the model may make mistakes in details like price and availability, and recommends visiting the merchant's site to confirm accurate information.

Key points

• ChatGPT launches Shopping Research, a feature that generates personalized shopping guides by researching products online
• Available since November 24 for Free, Go, Plus, and Pro plan users on web and mobile, with nearly unlimited use during holidays
• Asks clarifying questions about budget, recipient, and features before researching
• Uses memory from previous conversations and allows guiding the search in real time by marking preferences
• Powered by GPT-5 mini trained specifically for shopping tasks through reinforcement learning
• Works especially well in technical categories like electronics, home, kitchen, and sports
• Results are organic based on public sites and conversations are never shared with retailers
• OpenAI warns it may make mistakes in details like price and availability