Pixyle AI Raises $1 Million to Fix E-Commerce Product Search
Pixyle AI uses advanced computer vision to automatically tag and categorize fashion items, helping online retailers reduce cart abandonment. The startup just secures a €1 million seed round to expand its B2B visual search technology.
Pixyle AI secures €1 million (about $1.05 million USD) in a seed round led by South Central Ventures to tackle poor product discovery in online retail. Founded in 2019 by Svetlana Kordumova during her computer science PhD studies, the startup uses neural networks to analyze and tag fashion images just like a human would. The company currently works with over 20 clients, including major brands like Depop and Otrium, and successfully tags over 250 million images.
The technology targets the major issue of cart abandonment, which Google Cloud research shows affects 52% of shoppers who leave a site if they cannot find a specific item. Poor search results usually stem from incomplete or inaccurate product metadata provided by brands or individual sellers. While retailers traditionally hire humans to manually fix these data gaps, Pixyle AI automates the extraction of detailed attributes like color, pattern, and garment type directly from photos.
By automating this labeling process, Pixyle AI eliminates subjective human errors, such as one person calling a garment yellow while another calls it orange, which is especially crucial for massive secondhand marketplaces. The system allows shoppers to use highly specific queries, such as "short summer dress with flower print in pink and purple," and receive exact matches. As a result of implementing this visual AI, the startup reports an average conversion increase of 10% for its retail customers.