Ukrainian online stores are increasingly receiving customers not directly from people, but through artificial intelligence. According to Prom.ua, the number of conversions from AI services increased fivefold in 2025 and already accounts for about 1% of total traffic.
Ukrainian online stores are increasingly receiving customers not directly from people, but through artificial intelligence. According to Prom.ua, the number of conversions from AI services increased fivefold in 2025 and already accounts for about 1% of total traffic.
Similar dynamics are recorded at Epicenter: they say that traffic from AI has grown by about 500% per year, although it still remains at the level of single percentage points, reports LIGA.net. At the same time, the impact is much greater than clicks show: users often receive a recommendation in the chat, and then search for the product themselves.
Rozetka adds that up to 60% of users already interact with artificial intelligence in some way before making a purchase. Those who perceive shopping as a routine are the ones who move most quickly to this scenario, while for others the process remains more «manual.»
It’s about changing the consumption model itself: instead of searching and comparing dozens of tabs on their own, users are increasingly delegating their choices to AI assistants that take into account budget, preferences, and context.
Technically, this works through RAG approaches (resource generation with augmented sampling), where models like Perplexity, Microsoft Copilot, or Grok access external sources — search engines, directories, and databases — and generate recommendations based on them. At the same time, different services provide different answers: for example, Gemini has direct access to Google Search, while ChatGPT uses Bing-based search.
In response to these changes, Ukrainian retailers are already adapting their platforms. Prom.ua talks about transforming from a classic catalog into a data ecosystem and introducing formats that are understandable to language models. Rozetka is reviewing the structure of product cards and adding aggregated reviews, and Epicenter is experimenting with semantic tags and description logic.
In addition, businesses are beginning to prepare technical infrastructure for direct interaction with AI — in particular, APIs and MCP connectors through which assistants can directly connect to catalogs.
Experts note that classic SEO is already being supplemented by GEO (generative engine optimization) — optimization for generative search engines. In the new reality, it is important not just to be in the results, but to become a source that AI trusts.
«The business that wins is the one whose infrastructure can become the only source of truth for the AI agent,» Prom.ua summarizes.
