Using AI Technology to Mitigate Costly Returns

Pioneering uniform retailer and rental company, Uni-Mate, was looking to improve core business operations by harmonizing three key elements: product, technology, and operations. Monstarlab provided end-to-end service from planning up to the development of “AIxR Tailor” (Air Tailor) App. It is a Progressive Web App (PWA) [1] that enables automatic body size measurement using AI image recognition technology

Challenge

When it comes to client uniform production orders, sizes are normally provided by the customers based on their own untrained measurement efforts which often lead to size discrepancies. In effect, the business incurs increasing costs on returns and exchanges even reaching a staggering 40% return rate at the highest.

On top of that, the company has to carry more inventory than necessary to address size exchanges that often end up as wastage resulting in detrimental environmental impact.

This led Uni-Mate to aim for the development of an automatic measurement app that can accurately determine uniform sizes way before production leading to improved operations and customer service.

Solution

Uni-Mate initially wanted a native mobile app to be built, but the team proposed a PWA instead owing to a lot of benefits. A PWA supports multiple devices and does not require a separate download. It also greatly reduces the time and effort required for handling store applications and version support thus reducing development and operation costs.

AI specialists were then tapped for the development of the AI engine for generating accurate body size measurements. The AI team did careful research including engaging with veteran staff with years of experience in taking manual measurements.

The team also held regular collaboration with Uni-Mate on the development of the matching logic that would derive the appropriate clothing size for each person. Working together to create an AI engine that can support the process from defining requirements to designing details.

UI and UX components were also handled by Monstarlab through our design subsidiary, A.C.O. Crafting the tone and manner, designing the interface and illustrations that ensure usability and complement the required AI engine’s functionalities.

Results

Through technical research, a method of creating a 3D model from an image and predicting the actual size from it was derived and successfully led to the development of an original AI engine. The accuracy of the AI image recognition was improved through repeated verification using measurement data provided by Uni-Mate.

The application was built on a simple system that feeds back the appropriate size based on photos of the backside and profile of the person to be measured plus basic data such as height, age, weight, and gender. With UI design that took into consideration the large number of people to be photographed and managed, the developed app turned out to be an easy-to-use product.

Moving forward, Monstarlab will continue to conduct field tests and provide support to improve the value of service in terms of size determination, improvement of adaptation logic, and AI accuracy.

[1] What is a PWA?

Abbreviation for Progressive Web App, a new style of website that can provide a native app-like user experience on the web. Many advantages include no app installation required since it runs in the browser, adding an icon to the home screen, push notifications, faster-loading speed and display, and offline browsing.