Wayfair uses AI to automatically review, classify, and correct product attribute tags across its catalog of roughly 30 million items, covering details such as color, material, size, and style. This replaced a process that previously relied on suppliers and customers reporting errors, and on manual human review that could not keep pace with catalog volume.

Wayfair integrated OpenAI models into a tag-agnostic catalog enrichment system that classifies attributes across approximately 47,000 different product tags. A 'definition agent' ingests internal and web-based definitions to give contextual meaning to each tag; product data then feeds into a classification framework. As of March 2026, Wayfair had corrected 2.5 million product attribute tags across more than one million high-visibility products. The company also previously partnered with Snorkel AI and Google Cloud's Gemini to automate catalog tagging, reducing the time needed to curate product listings by 67% and improving some conversion rates by 2% according to those partnerships. When confidence in a correction is high, the AI updates the listing automatically and notifies the supplier; for uncertain or high-risk changes, it requests supplier confirmation first.