ZOZO's WEAR app adds AI outfit advice via 28 'fashion cores'

ZOZO is pushing its AI deeper into the fashion discovery process, this time inside WEAR, its app for sharing and browsing outfit posts. The new feature goes beyond the earlier "fashion genre" framing by adding what impression a wearer wants to give and how they want to style it, producing one of 28 so-called fashion cores per user. From there, the system sifts more than 15 million posted outfits to propose looks, and it can also account for body-shape concerns, the day's weather, and how to rotate a specific garment.

The feature leans on data ZOZO collected offline as well as online. From 2022 through 2025 the company ran a physical store in Tokyo called niaulab, where stylists made suggestions and talked with visitors, and those interactions, along with WEAR posts, were used to train the AI. Generative AI is used to explain styling points, and a face-photo upload lets users see a virtual try-on, with a path into ZOZO TOWN to buy.

The test for ZOZO is whether reducing the effort of finding clothes actually converts browsing into purchases, which is the stated aim. Much may hinge on how well the 28 fashion cores capture real preferences and whether virtual try-on images feel convincing enough to trust. Shoppers who dislike describing their own taste may find photo-based diagnosis easier, while ZOZO's store and app teams will be watching whether the suggestions drive purchases on ZOZO TOWN.