Slovenčina 
Čeština 
Shqiptare 
Österreich 
België 
български 
Bosanski 
Українська 
Eesti 
Македонски 
Türkçe 
Deutsch 
Dansk 
Español 
Français 
Suomi 
Ελληνικά 
Hrvatski 
Magyar 
Schweiz 
Italiano 
Lietuvių 
Latviešu 
Nederlands 
Norsk 
Polski 
Português 
Română 
Srpski 
Slovenščina 
Svenska 
Back in October, we covered how ChatGPT launched Instant Checkout, enabling direct purchases through the chat interface. Now OpenAI is taking another step – this time focusing on the research phase that comes before the actual purchase.
How It Works in Practice
The principle is simple – you write a request 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”. ChatGPT then asks clarifying questions about budget, preferences, and specific requirements.
The system subsequently searches the internet and checks prices, availability, reviews, and product specifications. It brings results progressively, and you can mark which products interest you and which don’t. After a few minutes, you receive a clear guide with top products, their key differences, and up-to-date information from reliable retailers.
Source: openai.com
Specially Trained Model
OpenAI deployed a modified version of GPT-5 mini for this feature, specifically trained for shopping tasks. The model learnt to read trustworthy sites, cite reliable sources, and synthesise information from multiple sources.
An interesting aspect is the integration with ChatGPT memory. If the system knows you’re a gaming enthusiast, it will factor that in when recommending a laptop. Professional users with Pro plans also get proactive suggestions through ChatGPT Pulse – for example, accessory recommendations if you previously discussed buying an e-bike.
Works Best For Complex Decisions
The feature excels in categories with numerous technical parameters – electronics, beauty, home and garden, kitchen appliances, and sports equipment. For simple questions like checking a price or verifying a feature, a standard ChatGPT response suffices. Shopping research makes sense for comparisons where you need to weigh multiple factors and trade-offs.
Transparency Instead Of Paid Recommendations
OpenAI emphasises that results are organic – they don’t come from paid partnerships with retailers. User conversations are not shared with merchants. The system reads publicly available e-shop pages, while low-quality and spammy sources are filtered out.
Merchants who want to ensure their products are included in results can go through an “allowlisting” process on OpenAI’s website.
Limitations And Accuracy
The company itself admits the model isn’t perfect. It can make mistakes with prices and product availability, so they recommend visiting the seller’s website directly for the most current information.
In internal tests, however, the new version achieves better results compared to previous models in accurately citing product details.
Availability
Shopping research is gradually rolling out on mobile and web for logged-in users of all plans – Free, Go, Plus, and Pro. When you ask a shopping-related question, ChatGPT automatically suggests using this feature, or you can launch it through the menu.
During the holiday period, the company has made nearly unlimited usage available for all account types, which is a significant expansion compared to standard limits.
