Recommendation Engine

A customer's product request is represented as a collection of attributes. Our recommendation engine allows for both strict and fuzzy matching against the shop's product database in order to find matching products. After applying filters based on the customer's preferences and collaborative filtering, a top selection of products is selected.

The recommendation engine is able to create whole outfits based on products available in your online shop. In cooperation with human stylists, we have created a generative algorithm which evaluates created outfits based on pre-defined and self-learned rules. Customer feedback is incorporated into the existing rules and applied for future iterations.

In addition to searching and filtering products through the chat, Inspora enables customers to find products similar to ones they like or already own. These products are ranked by matching attributes as well as their visual similarity based on product images.