1.27 Million Views Without Reshooting Every Combination
Ustyle's styling engine has logged more than 1.27 million outfit views. The core of that performance is not shooting new lookbooks continuously. It lies in combining a brand's existing product catalog per customer into complete, shoppable outfits made of items that can be bought immediately.
Italian AI startup Ustyle processes each product once and then applies actual product images onto an AI model. Showing every combination of tops, bottoms, and accessories on an online store normally requires a separate shoot for each combination, but this method assembles fitted looks from within the catalog. What shrinks is the burden of reshooting for every combination, and the result is a personalized, shoppable outfit.
An early customer, Turin-based leather brand Mialuis, attributed 33% of its total revenue during a 60-day pilot to this engine. Instead of the conventional 'new arrivals'-centered suggestions, it reflected context, styling logic, and user preferences, and was attached to the e-commerce site as a lightweight plugin. About a quarter of that revenue received an attribution score of 80 to 100. It is a case where recommendations were actually used in the purchase process while lowering the burden of shooting each combination.
