English
The technology behind all of this must handle enormous amounts of data, understand context, and take into account the behavior of each individual customer. Luigi’s Box has been working on this challenge for a long time. The GENAI-ECO project takes our research to the next level: we combine the latest advances in artificial intelligence, vector models, and behavioral data to build solutions that truly work in real-world environments:
1. Hybrid search that truly understands full-text search is fast, but doesn’t understand context
Vector models understand meaning, but can miss exact matches. The hybrid search we’re developing combines both approaches and adds an extra layer: re-ranking built on behavioral data and the business goals of each individual store. The result? A search engine that thinks like an experienced salesperson, not like a database.
2. An AI assistant that guides the customer to the right product
Sometimes a customer doesn’t know exactly what they’re looking for. Our AI Shopping Assistant helps them find products easier. It guides them through a conversation, refines their requirements, and displays relevant products directly from the store’s catalog. The entire deployment can be handled in two days.
What we measure
We track whether hybrid search genuinely performs better (goal: at least 8 successful A/B tests), and whether we can deploy the assistant quickly and repeatedly (goal: 7+ deployments within two days, 10+ pilot launches in total).
Why this matters to us
Better search doesn’t just mean better UX. It means higher conversion rates, greater average order value, and customers who come back. That’s a goal we share with every one of our clients.
This research is carried out as part of the project Application of GenAI for Improving Product Search and Discovery in E-Commerce (GENAI-ECO, project code: 17I04-04-V05-00077), co-funded by the European Union from the Slovak Republic’s Recovery and Resilience Plan. The project runs from October 2025 to June 2026.
