Skip to main content

How AI in Fashion Disrupts Resale, Returns and Fit Tech

After years of hype, artificial intelligence (AI) is taking hold in fashion, from optimizing recommerce experiences to helping shoppers find clothes that fit to reducing returns and personalizing the shopping journey.

Retailers are buying into AI technologies, both figuratively and literally. Coresight Research predicts that global retail spend on AI will reach $8.5 billion this year, up from $7 billion in 2022 and projects a 24 percent compound annual growth rate (CAGR) through 2030.

During a panel at Shopify New York earlier this month, Trove CEO Gayle Tait said that resale’s growth has come despite a significant barrier for the overall market—keeping garments in circulation relies heavily on data in a world where these products inherently don’t have any data attached.

A resale technology provider to Patagonia and Levi’s, Trove is currently investing in deploying computer vision models that help its brand clients match items back to their original digital record, and leverage data such as the sales history to better guide where to sell the product.

Trove also uses machine learning tech to group “like” items together so consumers aren’t inundated with too many choices. For example, computer vision helps Lululemon bring together the hundreds of off-catalog pre-worn tank tops it receives as returns and helps to sell them under a single SKU on its branded resale marketplace, Lululemon Like New.

Related Stories

“There’s a lot of black Lululemon leggings out there, and there’s different lengths and different sizes. We have the ability to be able to use computer vision models to then take those items, and break them down into SKUs that are then shoppable,” said Tait. “So crop leggings, pattern black leggings, leggings that have detail on the side—all of that is where we’re able to leverage machine learning to organize them. It’s impossible to take 200,000 items that are each a different condition…and size, and make them shoppable by human hand.”

Artificial intelligence is keeping the resold clothing out of the supply chain as well. Trove, which powers branded resale for Canada Goose, sees just a 5 percent to 10 percent return rate on the clothing sold in those marketplaces.

Computer vision is powering the body measuring capabilities of fit technology platform 3DLook in the company’s mission to make size charts obsolete.

To create someone’s 3D model, 3DLook uses computer vision-based algorithms to analyze 86 different points of measurement, according to Whitney Cathcart, the company’s co-founder and chief creative officer.

“The most exciting piece, frankly, for what we do is the impact that body data has on a brand, when you think about how you’re actually designing, creating patterns, and grading those patterns,” said Cathcart.

Cathcart said 3DLook can aid in reducing the returns problem as well, particularly bracketing—when someone buys a product in multiple sizes and plans to returns everything except the one that fits.

“If we can invoke a trust in a consumer that the size that we’re getting them is even close enough—if they get one garment in home, it’s such a pain in the neck to return that we have found that people will keep it,” Cathcart said.

During the panel, Bryan Amaral, founder and CEO at retail consulting firm Clientricity, shared his experiences from early in his career, saying that like many modern technologies, AI adoption largely requires outside factors to accelerate throughout the retail market.

When he worked with retailers including Saks Fifth Avenue and Brooks Brothers in the late 1990s, for example, both companies weren’t sold on the idea of body scanning technology.

“No retailer wanted to put a large booth on their floor,” Amaral said. “The problem was, you could put it in a flagship, but you couldn’t put it anywhere else. And it just was too expensive. It took up too much space. So some of the big problems in the early days [were that] you had a great idea, but it wasn’t a great enough idea to hit mainstream adoption.”

As mobile technology developed, making room for more accurate three-dimensional avatars and data collection capabilities, brands began to get comfortable with testing these types of technologies. Additionally, Amaral noted that the industry’s clothing to spec quality today “is much better than what it was 10-to-15 years ago, in most cases”, thanks to more precise laser cutting and mechanical cutting methods.

AI can aid many links along the supply chain, such as powering front-end retail experiences or the back-end technology, with Cathcart upbeat on the technology’s ability to scale.

“There’s a lot of deep learning, which is one of many different AI technologies, to spot fashion trends,” Cathcart said. “You think about 3D design and creating digital garments. And our solution enables brands to understand what the customer actually looks like, rather than just designing for somebody who you think might be your customer.”

Cathcart also said AI has implications for other areas within fashion, including manufacturing and robotics, style recommendations and even payments.

Fashion retailers that deploy AI-based technologies must learn to properly conduct “safe to fail” experiments, according to Tait.

“I think with any innovation, that’s a really good guiding principle, where can you do something with super-light lift to get data back. And the great thing about this is it is real time,” said Tait. “Fail fast and have a view and be open to getting it wrong and learning from those learnings just as much as you learn about the things that go right.”

When it comes to shopping, Amaral pointed out that consumers want to feel appreciated—and if that’s not there in the experience, they won’t think twice about shopping elsewhere.

“By leveraging the kinds of technologies that are available, we can now create those kinds of near-human personalized interactions in an online experience,” Amaral said. “If I can start to know who it is I’m interacting with, whether it’s in-store or online, and start to really give them relevant communication experiences and even change the way the website feels—the consumer will feel visible just by you communicating the things that are truly relevant to them.”