Fisia Group Brings AI Shopping to Nike Brazil With Doris.ai

Retailers struggle with high return rates and lost sales because customers face uncertainty choosing the right fit and style across digital and physical retail channels. This challenge is significant as it directly impacts profitability, operational efficiency, and competitiveness in an industry where seamless, personalized shopping experiences are increasingly expected. 

Doris.ai, in partnership with Marvik, is solving this challenge by delivering AI-powered virtual try-on allowing customers to virtually try on clothes with accuracy. Their solution empowers retailers to enhance the customer experience, streamline operations, and drive higher conversion rates through immersive, data-driven shopping journeys.

The boundary between digital and physical shopping is dissolving, ushering in a new era of experience-commerce, the bridge between in-store and online retail where technology and human experience merge. Doris.ai, part of NVIDIA’s independent software vendor (ISV) ecosystem, is redefining how customers engage with fashion, no matter where or how they choose to shop.   

Doris.ai’s approach is remarkably user-friendly: With just two photos from a shopper’s phone—one from the front and one from the side—the platform instantly captures accurate body measurements. This unlocks a sophisticated virtual try-on experience.  

Whether browsing online, interacting with a smart mirror in-store, or using a mobile device on the go, shoppers can see themselves in any outfit, mix and match styles, and experiment with different sizes on a highly realistic, personalized 3D representation of their own body in real time. The result is a seamless, deeply personal shopping journey that adapts to every context.

Behind the scenes, Doris.ai leverages NVIDIA® TensorRT™ and Triton Inference Server™ to deliver tailored recommendations and answer nuanced customer queries. The platform’s virtual try-on and avatar generation capabilities are powered by a combination of seven APIs, hosted on a Kubernetes cluster of AWS instances equipped with NVIDIA A10 Tensor Core GPUs. Each API can scale horizontally based on traffic, ensuring robust performance even at peak demand. 

This high-performance infrastructure enables Doris.ai to deliver a seamless, real-time experience for shoppers and retailers alike. Drawing from product catalogs, purchase histories, and real-time inventory data, the platform suggests outfits, recommends sizes, and highlights trending styles. Its multi-query capability allows customers to interact naturally—asking for fit advice, alternate colors, or complementary pieces in a single, fluid conversation. 

To develop the sophisticated technical capabilities that power its experience-commerce platform, Doris.ai partnered with Marvik, an AI consulting firm and service delivery partner (SDP) with deep expertise in computer vision, generative AI, and predictive analytics.  

Marvik developed and optimized generative models, such as Stable Diffusion, to enhance garment visualization, and built computer vision algorithms to accurately extract body measurements from user images. It also engineered robust recommendation systems that personalize garment suggestions based on both predicted fit and product attributes. By optimizing these solutions, Marvik enabled Doris.ai to deliver high-resolution, real-time performance at scale across both online and in-store environments.