ChatGPT launches conversational shopping research tool

A screenshot of ChatGPT results for winter shoe recommendations / Captured from ChatGPT

OpenAI has launched a new shopping research feature in ChatGPT, aiming to transform the way users explore and compare products before making purchasing decisions.

The feature is designed to create a more personalized and convenient shopping experience by enabling users to conversationally discuss products they are looking to buy directly with the artificial intelligence (AI) chatbot.

Regardless of their subscription status on the platform, users can access the feature by describing their needs. ChatGPT scans the web for product options, prices, reviews and related images.

Depending on the product, ChatGPT can also ask follow-up questions about budget, preferred features and styles to help refine its recommendations. As a result, users get a personalized buyer’s guide, including top picks, side-by-side comparisons and direct links to purchase items.

OpenAI said the feature respects user privacy, with no personal data shared with merchants, and that all results are generated organically from publicly accessible sources.

When testing it with a search for “good sneaker-type shoes for winter,” ChatGPT asked about budget and for clarification about the weather and circumstances to find appropriate products for the scenario.

After being given further instruction to look for shoes under 150,000 won ($101) to wear in light snow, it provided three suggestions, each with rationale for purchase and comments to help users match their specific winter needs.

A screenshot of ChatGPT results for winter shoe recommendations / Screen captured from ChatGPT

Microsoft’s Copilot gave similar but less refined answers. The product options included five picks from major brands, with brief descriptions, purchase links and budget-friendly alternatives from outlets and online platforms.

Google’s Gemini gave the least tailored result, giving only a basic list of major brand products categorized by three key features, with vague price ranges and no purchase links or details.

ChatGPT did the best job of providing tailored results by asking specific questions and offering detailed product guides, pros and cons, and tips for using the item.

It also offered a wider range of options, providing tailored choices and taking affordability, warmth or style into account.

While Copilot merely presented a straightforward product list with links, ChatGPT took the shopping experience further by offering up-to-date results that reflected the current inventory on major Korean online platforms like Coupang, Gmarket and 11st, enabling users to make decisions based on real-time availability.

However, it lacks integration with one of Korea’s largest shopping platforms, Naver Shopping, which continues to dominate the local e-commerce scene with a vast array of domestic brand vendors and millions of user reviews.

When searching for winter sneakers on Naver, the most-reviewed winter footwear product — Discovery Expedition’s padded shoes — was immediately featured, reflecting local trends and consumer feedback.

ChatGPT’s results, though comprehensive and up-to-date for platforms such as Coupang, did not surface these top-rated items — likely due to its limited access to Naver’s proprietary database and localized product information.

This highlights a key limitation of the new shopping tool from a Korean user perspective, as Naver remains the dominant shopping platform. For anyone seeking the most locally popular, community-endorsed items, this gap is significant.