US company uses AI to create virtual mannequins for hard-to-fit bodies

For years, sisters Margaret and Janice Tam struggled to find clothes that fit.

Self-described as petite, they said their shopping journeys often ended with settling for the item that best fit their shape or buying a garment only to alter it later.

Their frustrations sparked an idea to develop software that uses body-scanning technology and predictive analytics to create virtual mannequins, or avatars, and inform shoppers how an item would fit on certain parts of the body, such as their shoulders, waist or hips.

In just two years, Tam Technologies – doing business as TrueToForm – has built a client list that includes fashion studios and clothing brands in the US and Europe.

Their latest product is Fitsearch, a garment-focused search engine that provides a curated list of recommended places to shop based on body dimensions.

Using the camera function on cell phones or tablets, the company’s app creates a 3-D model silhouette of a person’s body from more than 60 measurements collected from uploaded images. Clothing brands embed the software onto their websites to remotely collect a customers’ measurements so shoppers know how certain garments will fit. Other brands use it for their design teams.

“We want to be like Kayak, but for clothing,” said Margaret Tam, the company’s chief technology officer. “As big as Kayak is in terms of searching and helping people find flights and compare different providers, that’s what we want to be for clothing.”

So far, among its more than 40 corporate users are the Tailory, a New York company that makes high-end custom clothing, Vermont-based winter gear company Burton and Swedish winter lifestyle brand Ridestore.

TrueToForm eventually will add a geolocation feature that, after walking into a clothing store, would automatically provide clothing recommendations based on a person’s 3D body dimensions, bypassing the need to search, Janice Tam said.

Sparked by frustration

In 2020 during pandemic lockdowns, Margaret Tam purchased a sewing machine to alter clothing items in her closet. She also searched online for dress forms, a foam mannequin tailors use to make patterns and measurements for garments, but couldn’t find any that matched her body shape.

“They’re all very much standard body proportions, like an hourglass figure,” Margaret Tam said. “It’s useless to get a dress form that doesn’t match your body because then all the clothes you make are not going to fit you. It’s going to be the same problem.”

Her initial idea was to start a manufacturing company that produced customized dress forms. To Janice, her economics-minded sister completing her master’s degree at the University of Chicago’s Booth School of Business, it didn’t seem like a profitable idea. The profit margins, if any, would be too low given it’s a niche product that is costly to manufacture and ship.

With consumers increasingly shopping online, Janice saw an opportunity to digitize the fitting room experience with 98% accuracy. Margaret, at the time a software engineer at Apple Inc., quit her job to commit to building the platform.

“We realized it was a really interesting time in the fashion industry because in the middle of Covid, everyone started using virtual design tools,” Margaret said. “Covid happened and suddenly all of these fashion brands started virtually prototyping and designing and draping all their clothes on virtual dress forms instead of using physical mannequins.”

After building the intellectual property and developing the software in 2021, the sisters launched a prototype in 2022, aided by a small sum from investors. In 2023, they added features using a US$50,000 (RM234,800) equity-free prize from MassChallenge, a startup accelerator in Boston. Earlier this year, they were awarded a US$35,000 (RM164,360) innovation grant from Minnesota to help grow the business.

The business relocated to Rochester in 2023 when Janice, the company’s chief executive, moved there with her husband. Margaret remains in California.

A solution for a growing problem

For about a decade, companies tried using large hardware body scanners inside malls and research institutions to measure various body shapes. It wasn’t scalable, said Janice. Other companies tried using survey data to provide retailers and their customers with size recommendations. For hard to fit body types, it wasn’t an efficient model, she said.

Advances in generative artificial intelligence and AI models changed that. “We’re able to take the accuracy of those hardware scanners and bring that to the convenience of a smartphone,” she said.

The company’s technology also addresses a costly problem for retailers. In 2023, retailers lost US$743bil (RM3.48 trillion) in sales due to returns, according to the National Retail Federation. For every US$1bil (RM4.6bil) in sales, the average retailer incurs US$145mil (RM680mil) in merchandise returns, NRF reported.

Statistics show the leading category of returned items are clothes, with the top reason being the items didn’t fit.

The problem presents a US$93bil (RM436bil) opportunity for TrueToForm, Janice said.

“Our goal is to have this become available to all shoppers, not just a single brand shoppers,” she said.

Janice doesn’t anticipate TrueToForm being profitable for some time, since all profits are being reinvested in company technology. The goal is to build the company’s value and make it attractive for acquisition.

“Ideally, our target acquirer would not necessarily be an individual brand but a large retailer or a big tech company like a Google or Microsoft … that could really take advantage of our search engine as well as our ecosystem of both brands and shoppers,” Janice said.

“We want to be a one-stop shop for searching and discovering clothes that would fit you,” Margaret said. – Star Tribune/Tribune News Service

Tagged