wisello

wisello lets a shopper describe what they need in their own words. The assistant searches the store’s catalog, asks clarifying questions, and recommends products in the language the shopper is using.

It began as my final thesis project at Universidad ORT and launched as a product in 2023. The thesis received first place from Uruguay’s Academia Nacional de Ingeniería.

Shopping as a conversation

A shopper can start with a plain-language description, ask about a product, or look for alternatives. The conversation provides a way to explore the catalog without having to know the store’s exact product names or categories.

The assistant supports questions in different languages, keeping that interaction inside the shopping experience.

Connecting the catalog

wisello could connect to a store through CSV, XML or JSON feeds, an API, or an e-commerce platform integration. The catalog supplied the products for the assistant to search and recommend.

The front end used React and Next.js, with an LLM-based conversation layer over the store’s catalog.

From thesis to product

The project brought together the conversational interface and the catalog integrations needed to offer it to online stores. It moved from a university thesis into a launched product; it is now retired.