Fanatics uses AI to suggest matching team gear on every product page

Fanatics, the sports merchandise giant, has added a feature called Shop the Look. On each product page, it suggests other in-stock items that go with the one a shopper is viewing, such as a hat and socks to match a jersey. The company built it with help from OpenAI's Codex, and says views of recommended items rose 10% after launch.

What changed

Shop the Look picks items based on team and intended style. Fanatics says it uses a tool built with Codex, OpenAI's coding product. The tool reads product and catalog data, finds items that fit together, and rechecks its own recommendations. The result is a ready-made outfit of team gear built around the item a fan is considering.

Why Fanatics needed it

The catalog is huge. Fanatics offers more than 1.5 million unique items at any moment, and releases upwards of a thousand new products each week. Justin Tsai, chief product officer for Fanatics Commerce, said shopping sessions often run over 30 minutes on average. Customers are searching a very large catalog to figure out what they can buy.

Photo shoots and styled outfits are the traditional fix. Tsai said they take too long for a catalog this big, and for fast items like championship shirts and hats.

Why team matching matters

Random suggestions would backfire with sports fans. Tsai put it this way: nobody who is a Los Angeles Dodgers fan wants to be matched with a San Francisco Giants item. Keeping suggestions on the right team and player is central to the feature.

The bigger picture

Team merchandise is still Fanatics' biggest source of revenue. CEO Michael Rubin said its commerce business will top $7 billion this year, against roughly $5.5 billion for collectibles. Fanatics also makes more than 45 million items a year and sells to customers in more than 180 countries.

For OpenAI, this is one more shopping effort. CNBC reports it is ending Instant Checkout, which let people buy directly inside ChatGPT. It is now working with retailers on dedicated apps within ChatGPT.

What is not known yet

The 10% figure measures views of recommended items, not sales. The article does not say whether shoppers bought more, or how many Fanatics sites now have the feature. Tsai did say the suggestions give shoppers more conviction about a purchase by showing how it fits with other team gear.