Wildberries launches AI product comparison for its 100+ million users

The Russian marketplace Wildberries has launched an AI-powered product comparison feature. The neural network analyzes several products a user has selected and determines which one is better and why, in order to simplify the choice. For now the option is not available to all users and runs in test mode, but the company promises to open it to a wider audience in the coming days.

How AI comparison works

According to CNews, the new feature is built on analysis of the product pages a user picks for comparison. The neural network compares their specifications and uses them to formulate a conclusion about which option is preferable and what exactly its advantage is. This sets Wildberries' approach apart from the simple spec table that has long existed on many marketplaces: instead of making shoppers match dozens of parameters in a table themselves, the system immediately gives a substantive conclusion in plain language.

In recent years, other large online retailers outside Russia have also launched similar projects that bring generative AI into the customer experience, embedding conversational and comparison assistants directly into product pages or search results. The logic is the same everywhere: the more complex the assortment and the more options a shopper has, the more valuable a tool that handles the initial analysis and shortens the time needed to decide on a purchase.

Similar logic is already used in marketplace recommendation systems, but there the algorithm usually just ranks products by relevance to the query, without explaining the reason for the choice to the user. The comparison feature goes further: it formulates a substantive argument for a specific option, which is closer to how a live sales assistant works and could increase the shopper's trust in the recommendation, not just in the product's position in the results.

• Company: Wildberries, one of Russia's largest marketplaces.
• Platform audience: estimated at more than 100 million users.
• Feature: AI product comparison.
• Launch status: test mode, available to a limited number of users.
• Plans: expanding access in the coming days.

Why this matters at this scale

On a platform with an audience measured in hundreds of millions of visits, even a small conversion increase from easier product selection translates into a noticeable economic effect for the marketplace. The problem of choice on marketplaces is well known: the same product is often sold by dozens of different sellers with different prices, bundles and reviews, and the specs on product pages are not always directly comparable. A tool that takes over this analytical work and delivers a clear recommendation reduces so-called "choice fatigue", one of the main reasons users put off a purchase or leave the site without making one.

For Wildberries, launching such a feature is also part of the competition with other large Russian marketplaces, which are likewise investing actively in AI tools for shoppers. As the assortments and prices of the big platforms increasingly overlap, service quality, including how easily and quickly a shopper finds a suitable product, becomes one of the few remaining areas for differentiation.

What this changes for shoppers

For the shopper, such a tool effectively turns product search into a consultation, similar to talking with a sales assistant in an offline store, only without time limits and without pressure from the seller. It is part of a broader e-commerce trend in which large platforms worldwide are building AI assistants into the buying process itself, from picking a product based on a description of the need to comparing alternatives and explaining why a particular option fits better.

The test status of the launch suggests the company is still assessing how users react to AI recommendations and how much they trust this comparison format before opening the feature to the platform's entire audience. The success of such features is usually measured not only by conversion growth but also by fewer returns: if the AI helps a shopper pick a truly suitable product, some of the returns caused by wrong choices disappear on their own.

Rolling out such a feature to the whole audience usually happens in stages. First the accuracy of the recommendations is evaluated, along with how closely the option suggested by the AI matches the shopper's subsequent choice. Only then is the feature scaled to tens of millions of users, without the risk of making an outright bad recommendation across the entire platform.