Shopping online has become easier in one sense and harder in another. Consumers can access thousands of products in seconds, but choosing between them can still mean opening multiple tabs, comparing specifications, checking reviews, and deciding whether a “deal” is actually a good one.
PriceHub.AI is being built around a different idea: instead of forcing shoppers to translate their needs into filters, let them describe what they want in normal language and let the system do the matching.
The project comes from the team behind E-Katalog, a product comparison ecosystem with more than 20 years of experience structuring product information, retailer offers and availability. The company already operates services for different markets, including E-Katalog in Ukraine and E-Catalog in the United Kingdom. PriceHub.AI is being developed as a separate AI-powered service on top of that data infrastructure.
From a request to an actual product
A shopper might ask for a lightweight laptop for video editing while traveling, or a robot vacuum for a particular home and budget. The system first has to understand the use case, identify which specifications matter, match suitable products and check what is actually available.
One of the harder problems is product matching. Retailers often describe the same device differently, use different product codes, or bundle different accessories and configurations. Before comparing offers, PriceHub.AI has to determine whether two listings really refer to the same product.
The system can then compare retailer offers, show alternatives, summarize user feedback and examine how a product’s price has changed over time. In the beta version shown to Ekonomichna Pravda, the service could take a broad request and turn it into a specific recommendation while keeping follow-up questions inside the same conversation.
AI is the interface, not the database
This is where PriceHub.AI differs from a general-purpose chatbot.
The team does not position the language model as the source of current market knowledge. Specialized modules gather and match relevant information first, and the language model turns those results into a readable response.
That distinction matters for shopping. A model may explain the difference between two technologies, but a useful buying assistant also needs to know whether a specific configuration is available locally, whether two stores are selling the same version, and how the current offer compares with the broader market.
Andrii Bashlak, Head of UX/UI & Product Experience at PriceHub.AI, described the goal as helping shoppers understand what fits their needs, where they can buy it and whether now is the right time — without making them work through the underlying data themselves.
PriceHub.AI can also use price history to add context to a recommendation. Instead of only showing what a product costs today, the system can examine how its market position has changed over time. In tests demonstrated to Ekonomichna Pravda, it also summarized large numbers of customer reviews into a more concise signal that could help users assess owner satisfaction.
European backing for the technology
PriceHub.AI has also received external recognition. In 2026, the project was selected through a European Innovation Council program supporting Ukrainian technology companies. According to Ekonomichna Pravda, it received approximately €500,000 to develop its technology core, local models and orchestration system, and to prepare the product for market testing.
The first full-scale test is planned for Ukraine, where E-Katalog already has an established audience and retailer relationships. The team then plans to expand to Poland and the United Kingdom, followed by the United States and other markets.
The bigger idea behind PriceHub.AI is not simply to add another chatbot to online shopping. It is to change the role of a comparison platform itself. Instead of making users search through specifications and store listings until they reach a conclusion, the system is designed to start with the shopper’s real-world need and work backward toward a specific, explainable choice.
If it works as intended, the useful part of AI shopping may be less about generating more information and more about reducing the amount of information a consumer has to process.
This content is provided for informational purposes only and is not a substitute for professional advice. AFP editorial staff were not involved in the creation of this content.
Share
