YandexGPT has learned to summarize product reviews

Yandex Search users will now be able to quickly identify important things to consider when shopping online, and they will be informed of commonly mentioned pros and cons regarding their purchases. To do this, the YandexGPT neural network summarizes customer reviews into concise lists. These summarized reviews are also available on Yandex Market.

The results appear in product cards on both platforms. Users can evaluate the list of pros and cons generated by the neural network, and, if there are any inaccuracies, they can promptly report them with the press of a button. In Yandex Search, these lists appear when specific model names are searched for, such as "iPhone 14 Pro" or "Yandex Station Mini." Clicking on each advantage or disadvantage opens up reviews in which that particular feature is mentioned.

Example of YandexGPT product reviews highlight (the original version is in Russian only)

YandexGPT analyzes reviews left by customers on Yandex Market and in the "My Reviews" section of Yandex Search. The algorithm initially selects detailed reviews with a description of user experience. The neural network then highlights the product features that people consider important and mention most frequently. To create a list of pros and cons using YandexGPT, a minimum of 10 high-quality reviews per product is required. YandexGPT replaces the smart review generation technology introduced in 2021 on Yandex Market.

Yandex Search has the largest database of products in the Russian segment of internet, with nearly a billion listings from 60,000 stores and marketplaces. Users can find and compare products, access detailed information about them, learn about pros and cons, find competitive prices, and navigate to stores once they've decided to buy.

Yandex Market offers over 53 million products, helping users order everything they need. With almost 62,000 partner stores, the marketplace can handle order processing, delivery, and customer service.
 

Contacts:

Press Office
Polina Pestova
E-mail: pr@yandex-team.com