Home goods retailer Wayfair has deployed OpenAI models across its operations to enhance supplier support and product catalog quality at scale. What began as small-scale experiments in 2024 has evolved into a full production system that reduces manual effort and improves data quality across 30 million products.

Wayfair's catalog team faced the challenge of managing tens of millions of products across nearly a thousand classes, requiring consistent attribute tags for search and recommendations. The company built a reusable AI architecture using a single OpenAI model with a "definition agent" that produces contextual meaning for each of 47,000 tags. This system has processed over 1 million products and is expanding model coverage 70x faster than previous approaches.

The company also integrated AI into supplier support workflows through Wilma, an agentic platform that automates ticket triage and resolution. The system reads incoming requests, fills in context, and routes issues to appropriate teams. Wayfair reports automating 41,000 tickets monthly and reducing turnaround times significantly. Teams track "alignment rate" between AI recommendations and human decisions, gradually shifting from assistive to semi-autonomous modes as trust builds.