What it does

E-shoppers often spend hours going through overwhelmed images to find one piece to order. The process of browsing, choosing, ordering, trying on and returning is time-consuming and frustrating for many consumers and retail stores. Many returned products end up going in landfills, rather than going back to the store. Each year, 5 billion pounds of waste and 4.7 million metric tons of CO2 are generated through returns.

NONA provides an efficient, accurate, and simple way for fashion brands and retailers to implement online 3D avatar models. Online customers can use NONA to build personalized 3D avatars based on their conversations with an AI-powered chatbot and self-measurement. Digitizing a customer includes generating an avatar of the human body that resembles the customer. A perfect 3D avatar will match the customer in terms of height, face, body size, shape, dimensions, etc. This 3D avatar will be saved on our platform and can be viewed on their phone. It significantly decreases product return rate by offering a more accurate and realistic size and fit on the 3D avatars. More than having one size, customers can have multiple sizes and shapes based on the measurement. NONA can leverage the power of analytics to reduce anxiety for shoppers by offering highly-personalized style recommendations based on past selections, shopping history, and browsing patterns to promptly select and recommend and assist customers in expanding their dressing style efficiently.

Online shoppers will enjoy a better buying experience, leading to a significant drop in the number of returns and landfill waste. As a result, sales and profit will increase for retailers. Though the implementation of our 3D avatar platform, the personalized virtual model is bound to revolutionize the retail industry and enable fashion sustainability.

How we built it

Challenges we ran into


Machine learning model for predicting fit was only 60%

Accomplishments that we're proud of

What we learned

What's next for NONA

Augmented Reality

Share this project: