Lenta tech has developed the AI Platform Trends tool to analyze market demand across Ozon, Wildberries, and Yandex Market. The system aggregates data on sales, search queries, ratings, and reviews to identify trending products. It uses a hybrid approach combining rules, machine learning, and LLMs to semantically match and merge duplicate product listings from different platforms into a single entity.
The platform categorizes products and tracks their lifecycle across five phases: inception, growth, seasonal growth, explosive growth, and decline. Users can filter data by sales volume, revenue, and search interest, with the ability to compare performance year-over-year or period-over-period. This allows managers to distinguish between temporary spikes and sustainable trends.
The tool features a SKU verification mode that processes up to 100 items per batch via XLS files. For products already in the Lenta assortment, the system provides competitive price analysis. This functionality helps the commercial team make data-driven decisions on whether to test, expand, or discontinue specific items.
Since its implementation, the platform has reduced category analysis time from several hours to under 10 minutes. Currently, more than 200 employees use the tool. It has facilitated the selection and listing of over 150 trending SKUs, contributing to a 10% increase in retail turnover (RTO).
Future development plans include expanding data collection to international marketplaces and refining the semantic matching process. The team aims to improve the detection of early-stage trends to identify high-potential products before they become widely popular.
