OpenAI ChatGPT Shopping Research Tool Asks You Questions First — Putting Shoppers In Control

OpenAI has launched a new 'shopping research' experience within ChatGPT that flips the traditional online shopping script, asking clarifying questions before delivering personalised buying guides rather than immediately generating product lists based on a single query.

The feature, which rolled out on Monday across mobile and web for all logged-in ChatGPT users, transforms product discovery into a conversation. Instead of sifting through dozens of retail sites, shoppers describe what they need in plain language, and ChatGPT builds a customised buyer's guide by asking smart questions, researching online, and incorporating past conversations to understand preferences.

How It Works

The shopping research tool operates through a dedicated visual interface accessed either automatically when ChatGPT detects a shopping question or manually via the plus menu. Users might type requests like 'Find the quietest cordless stick vacuum for a small apartment', 'Help me choose between these three bikes', or 'I need a gift for my four-year-old niece who loves art'.

Rather than immediately generating results, ChatGPT opens a quiz-like experience to gather more context. It asks about budget constraints, colour preferences, size requirements, must-have features, and who the item is for. As users answer these questions, the tool refines its understanding and begins pulling relevant products from trusted sources across the web.

During the research process, ChatGPT presents 10 to 15 potential items and prompts users to swipe right on products they like or left on those they don't, similar to dating apps. This interactive feedback loop helps the AI narrow down options in real time, adjusting to preferences as they emerge.

Within a few minutes, users receive a detailed buyer's guide featuring product comparisons with pros, cons, key tradeoffs, current prices, specifications, availability, and reviews sourced from what OpenAI considers high-quality websites. If memory is enabled, ChatGPT can draw on information from previous conversations to further personalise recommendations.

The Technology Behind the Tool

Shopping research runs on a specially trained version of GPT-5 mini, a model trained via reinforcement learning for shopping tasks. OpenAI trained it to read trusted sites, cite reliable sources, and synthesise information across many sources to produce comprehensive product research.

The company says user experiences shared on platforms like Reddit may be considered more trustworthy than paid marketing or reviews posted directly on product pages. However, teaching the model to identify objective, unpaid, or genuine user reviews has proven challenging.

The tool performs exceptionally well in detail-heavy categories like electronics, beauty, home and garden, kitchen appliances, and sports and outdoor equipment. For simple shopping questions such as checking prices or confirming specifications, OpenAI recommends sticking with regular ChatGPT responses, which deliver faster answers.

Amazon's Conspicuous Absence

One significant limitation has emerged. Shopping research only surfaces information from websites that permit OpenAI's browsing agents to access their pages. When asked whether that includes Amazon, OpenAI clarified that the system follows each site's rules and respects robots.txt files.

Amazon has updated its robots.txt file in recent months to block several OpenAI crawlers, including agents used for model training, real-time browsing, and its SearchGPT engine. This means Amazon product listings may be limited or entirely absent from ChatGPT's shopping research experience.