Fanatics is using OpenAI's Codex to build an AI outfit-matching tool for sports fans

Fanatics is using OpenAI's Codex to power "Shop the Look," a feature that suggests coordinated outfits on product pages for sports fans, according to CNBC.

The tool pairs each item with other products that are in stock, taking into account the shopper's team and the style they have in mind, according to CNBC. It analyzes information about products and the catalog to identify items that go together, then reassesses its suggestions. A shopper considering a jersey might be shown a cap, socks and various other accessories next to it.

Justin Tsai, chief product officer for Fanatics Commerce, said views of recommended items rose 10% after the rollout, according to CNBC. He said the suggestions give shoppers "more conviction" about what they plan to buy by showing how an item fits with other team gear.

Fanatics sells over 1.5 million distinct products, or stock keeping units, across its sites. Fanatics Commerce makes over 45 million items each year and sells to buyers in more than 180 countries.

Tsai said the company operates hundreds of sites and that customers struggle to search its catalog. "They're always trying to look through our very large catalog to figure out what they can buy," he said, according to CNBC. Shopping sessions averaged more than 30 minutes, and Fanatics debuts 1,000 or more new items every week, he said.