ChatGPT has launched a new functionality designed to accompany users in their purchasing processes. This is “Shopping Research”, an integrated tool that allows you to generate personalized guides to compare products, understand differences, refine needs and make purchasing decisions with greater judgment and context.
In the middle of the shopping season, with Black Friday active and Christmas approaching, this is a strategic bet by OpenAI, which seeks to capitalize on one of the moments with the highest volume of commercial searches of the year.
The experience transforms the purchasing process into a guided conversation
“Shopping Research” is now available in web and mobile versions for all users, both in free and paid plans. The experience transforms the purchasing process into a guided conversation, in which the model asks relevant questions, researches trusted sources on the internet and generates a complete report with recommendations adapted to the user’s profile.
Unlike a simple price or feature search, the new functionality is optimized for complex decision cases, such as finding a powerful gaming laptop for under $1,000, comparing baby strollers based on terrain, or finding a creative gift for a 4-year-old girl who loves art. The tool acts as a personal assistant, asking segmentation questions (price, use, desired functionalities), displaying detailed sheets with images and links to distributors, and allowing the user to mark options such as “I’m not interested” or “More like this” to refine the results in real time.
The engine powering this functionality is a specifically tuned version of GPT-5 mini, trained by reinforcement for purchasing tasks. According to OpenAI, this model significantly improves accuracy in identifying products that meet user criteria. In internal tests, it has achieved a 64% success rate compared to 52% for the previous version.
Additionally, “Shopping Research” is designed to work in sync with other ChatGPT capabilities. If the user has memory activated, the system can recover previous conversations to refine its recommendations. For example, if someone has inquired about electric bikes, they might later receive guidance on relevant accessories.
As for the transaction, in the current version the suggested products redirect to the sellers’ official sites, but going forward, OpenAI plans to allow transactions to be completed without leaving the ChatGPT interface with the direct purchase feature called “Instant Checkout”.
In this context, artificial intelligence is positioned as an actor with the real capacity to prescribe products, influence purchasing decisions and transform digital commerce. And OpenAI’s proposal joins that of Google, which has also launched new purchasing functionalities in its “AI Mode”, and other agents such as Perplexity have explored similar experiences with its Comet browser. However, OpenAI’s proposal is differentiated by its level of interaction, customization and because, at the moment, the results are not sponsored.
Marketing professionals now have the possibility of influencing recommendations through presence in reliable sources and optimization of product sheets, opening a new line of work at the intersection between SEO, conversational experience and conversion funnel.
