Wildberries Launches AI Product Comparison for 100+ Million Users

Russian marketplace Wildberries launched a product comparison feature using artificial intelligence — a neural network analyzes several products selected by the user and determines which one is better and why, to simplify the selection process. For now, the option is available to not all users and works in test mode, but the company promises to open it to a wider audience within days.

How AI-powered comparison works

According to CNews, the new feature is built on analyzing product cards that the user selects for comparison — the neural network compares their characteristics and on this basis formulates a conclusion about which option is preferable and what exactly is its advantage. This distinguishes Wildberries' approach from simple tabular comparison of characteristics, which has long existed on many marketplaces: instead of making buyers independently compare dozens of parameters in a table, the system immediately produces a substantive conclusion in understandable language.

Similar projects to implement generative AI in customer experience in recent years have been launched by other major online retailers outside of Russia, embedding dialogue and comparison assistants directly in product cards or search results. The general logic is everywhere the same: the more complex the assortment and the more options a buyer has, the higher the value of a tool that takes on initial analysis and reduces the time to make a purchase decision.

A similar logic is already applied in recommendation systems of marketplaces, but there the algorithm usually just ranks products by relevance to the query, without explaining to the user the reason for the choice. The comparison function goes further — it formulates a substantive argument in favor of a specific option, which is closer to how a live consultant works, and potentially increases customer trust in the recommendation, rather than just to the position of the product in the search 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 — expansion of access within days.

Why this matters for an audience of this scale

On a platform with an audience measured in hundreds of millions of visits, even a small increase in conversion due to simplification of product selection converts into a noticeable economic effect for the platform. The problem of choice on marketplaces is well known: the same product is often sold by dozens of different sellers with different prices, product configurations and reviews, and the characteristics in cards are not always directly comparable. A tool that takes on this analytical work and formulates an understandable recommendation reduces so-called "choice fatigue" — one of the main factors due to which users postpone a purchase or leave the site without it.

For Wildberries, launching such a feature is also an element of competitive struggle with other major Russian marketplaces that are also actively investing in AI tools for customers. In a situation where assortment and prices of major platforms increasingly overlap, service quality — including how easily and quickly a buyer finds a suitable product — becomes one of the few remaining fields for differentiation.

What this changes for buyers

For a buyer, such a tool essentially transforms product search into a consultation, similar to a conversation with a sales consultant in an offline store, only without time constraints and without pressure from the seller. This is part of a broader trend in e-commerce, where major platforms around the world are embedding AI assistants directly into the purchase process — from product selection by description of need to comparison of alternatives and explanation of why a particular option fits better.

The test status of the launch suggests that the company is still evaluating how users respond to AI recommendations and how much they trust such a format of comparison, before opening the feature to the entire platform audience. The success of such features is usually measured not only by increased conversion, but also by reduced returns — if AI helps choose a truly suitable product, some returns caused by selection error disappear by themselves.

Expansion of such a feature to the entire audience usually occurs in stages — first the accuracy of recommendations and whether the AI-proposed option truly matches the user's subsequent choice are assessed, and only after that the feature is scaled to tens of millions of users without the risk of offering a clearly unsuccessful recommendation on the scale of the entire platform.