Fanatics rolled out Shop the Look, an AI-powered recommendation feature built with OpenAI Codex to suggest coordinated, in-stock outfits on product pages across Fanatics and official league stores.

The system uses enriched product, team, and inventory data and follows merchandising rules to ensure team-fit in recommendations, preventing mismatches like Dodgers items pairing with Giants gear.

OpenAI Codex analyzes catalog data to identify compatible items and reevaluate recommendations, creating cohesive looks such as a jersey with a hat and socks.

Early results show about a 10% lift in views of recommended items after the rollout, with shoppers gaining more confidence in purchases.

Testing across roughly 30,000 MLB products yielded a similar 10% increase in engagement for recommended items, signaling stronger shopper conviction.

Initial data also show a 10% rise in views of suggested items, as Fanatics Commerce handles over 45 million items annually and serves buyers in more than 180 countries.

Fanatics Chief Product Officer Justin Tsai says even small improvements in discovery can meaningfully impact conversion on a catalog that grows by thousands of items weekly.

CEO Michael Rubin told CNBC that Fanatics Commerce revenue is expected to surpass $7 billion this year, outpacing the roughly $5.5 billion from collectibles.

Rubin also noted the commerce business will exceed $7 billion in revenue this year, underscoring the scale of Fanatics’ commerce operations.

The AI initiative sits alongside existing strategies like curated photo shoots and in-person try-ons, with questions about how far AI will extend across the rapidly expanding catalog, which adds about 1,000 new products weekly.

The AI system was trained and tested with OpenAI technology, starting with 50 products in under a day and expanding to about 30,000 MLB products in broader testing.

Fanatics’ catalog exceeds 1.5 million unique SKUs, posing browsing challenges that the new tool aims to streamline by improving discovery across hundreds of sites and frequent product releases.