ChatGPT Introduces Shopping Research: AI Builds Personalised Buyer Guides

OpenAI has added a research-style shopping assistant to ChatGPT that promises to take the grunt work out of buying decisions. Called Shopping Research, the feature conducts live web research, asks follow-up questions to refine constraints, and assembles a short, evidence-backed buyer’s guide, effectively turning product discovery into a structured conversation.

Rather than returning a single quick answer, Shopping Research treats purchases as a multi-step problem: it probes user priorities (budget, use case, size constraints), scans high-quality retailer and review sources for up-to-date pricing and availability, and produces a side-by-side assessment of recommended options. OpenAI is initially making the tool available to logged-in users across Free, Go, Plus, and Pro plans, with near-unlimited usage for the holiday season to boost trial and adoption.

How Shopping Research works

Shopping Research opens as an interactive flow inside ChatGPT. Users describe what they want, for example, “Find the quietest cordless stick vacuum for a small apartment,” and the tool asks clarifying questions where needed. Under the hood, OpenAI says the experience is powered by a shopping-tuned variant of GPT-5 mini that’s optimised to read and synthesise product pages, reviews, and retailer information.

As the conversation progresses, the system updates results in real time. Users can mark options “Not interested” or request more items “like this”, steering the search dynamically. At the end of the session, ChatGPT returns a concise buyer’s guide with the top picks, key trade-offs, and the sources used, enabling readers to validate claims by visiting the original merchant or review pages.

Where it fits in the purchase journey

OpenAI positions Shopping Research for mid- to high-involvement purchases where trade-offs matter: electronics, appliances, personal care, sports gear, and other categories that require careful specification matching. For quick lookups, price checks, or single-feature confirmation, the standard ChatGPT reply remains faster and sufficient. Shopping Research targets users who want a curated, evidence-backed decision in minutes rather than hours.

Product accuracy and limitations

OpenAI includes internal benchmarking that places Shopping Research ahead of some baseline models on “product accuracy”, a metric that measures whether returned products meet user constraints (price, specs, etc.). Still, the company cautions that the tool is not infallible: price and availability are volatile, and the model can make mistakes. OpenAI recommends users verify final details on merchant sites, a necessary step given the real-time nature of retail data.

Privacy, transparency, and merchant participation

OpenAI says chats are not shared with retailers and that shopping research relies on publicly available retail sources. For merchants wanting visibility, OpenAI provides an allow-listing process so sellers can ensure product pages are discoverable by the system. The company emphasises transparent sourcing and cites as part of the user guide where data originated — a move that helps reporters and shoppers validate recommendations.

Business and platform implications

For OpenAI, Shopping Research extends ChatGPT deeper into e-commerce and consumer workflows, creating another touchpoint for users during high-traffic seasonal buying cycles. Allowing near-unlimited usage through the holidays is a tactical push to drive engagement and gather product-level feedback at scale. For retailers and platforms, the feature creates both opportunity and pressure: visibility in research results can influence sales, while inaccuracies or bias could raise reputational risks.

Risks for consumers and merchants

Two practical risks stand out. First, product pages and inventories change quickly; a recommendation that’s accurate at the time of generation may be outdated minutes later. Second, brands and sellers without allowlisting or clear metadata may be under-represented, skewing results toward well-indexed or large retailers. OpenAI’s allowance for merchants to follow an allow-listing process mitigates some exposure, but it leaves smaller sellers dependent on organic discovery.

Adoption metrics and quality indicators will determine whether shopping research becomes a mainstream buying tool or a niche convenience. Key items to monitor include the accuracy rate for complex filters (battery life, compatibility), user conversion from guide to purchase, and merchant uptake of allow-listing. Regulatory and transparency scrutiny could also follow if shopping research begins to materially shift traffic away from traditional comparison sites.

Shopping Research reorients product discovery from single-query lookups to guided decisions. For technology buyers and enterprise procurement teams, the tool offers a quick way to narrow choices and surface trade-offs; for consumer tech reporters and e-commerce operators, it’s a new aggregator to test for accuracy and market impact. As with any automated recommendation engine, final validation on merchant pages and a human check remain essential.