Walmart launches AI-powered virtual clothing try-on tech for online shoppers – TechCrunch

Walmart launches AI-powered virtual clothing try-on tech for online shoppers – TechCrunch

Last May, Walmart announced the acquisition of the virtual clothing try-on startup Zeekit, which leveraged a combination of real-time image processing, computer vision, deep learning and other artificial intelligence technologies to show shoppers what they would look like in an item through a simulation that takes into account body dimensions, fit, size, and even the fabric of the garment itself. Today Wal-Mart said it brings this technology to and its Walmart mobile app.

Retailer introduces ‘Choose My Style’ trial feature powered by a computer vision neural network, now in beta, which will now allow Walmart customers to select a style that better matches their appearance and body type body. At launch, online shoppers will be able to choose from 50 different designs to find the one that best reflects their skin tone, size and body type so they can get a better idea of ​​how the clothes will look on them.

These virtual models currently range in height between 5’2″ and 6’0″ and in sizes XS to XXXL. Walmart notes that it will continue to expand its model selection over time, with plans to launch nearly 70 more models in the coming weeks to provide more variety in height, skin tone and even skin color. hair.

The new feature is powered by Walmart Global Tech’s neural networks, which analyze clothing catalog images to create a dressed image using computer vision algorithms on a diverse set of Walmart model images, explains the company in its announcement on product launch.

The use of neural networks helps to determine the different variations available in a single product, which does not only include its size, but can also include other factors, like the color options available or even the length of the sleeves. . The new system is able to capture all of these variations when displaying options, the retailer explains.

Picture credits: walmart

The “Choose My Style” feature, which focuses on women’s fashion, will work with thousands of items from Walmart’s exclusive and private label portfolio, including Free Assembly, Scoop, Sofia Jeans by Sofia Vergara, ELOQUII Elements, Time and Tru, Athletic Works, Terra & Sky, No Boundaries, Avia and The Pioneer Woman. It is also adding national brands, starting with Levi’s, Hanes, Just My Size, Secret Treasures and Como Blu.

Over time, it will expand to include more national brands, including those from and Walmart Marketplace. The company notes that it is easier to activate the technology on its own brands because it owns the inventory and catalog of items.

Prior to its acquisition, Zeekit had worked with a number of retailers including Walmart, as well as other well-known brands like Tommy Hilfiger and Adidas. This existing work with Walmart accelerated the time it took to bring the technology to the public through Walmart’s website and mobile app integration.

Picture credits: walmart

When shopping on or in the Walmart app, customers will now see prompts that say “Edit my style” to select a style on the garment page if new virtual try-on technology is available for that product. These prompts will appear on as well as iOS and Android mobile apps.

According to Desi Gosby, Vice President Emerging Tech at Walmart, the model selected by a customer will continue to be presented with Zeekit-compatible garments for future shopping visits. However, the customer will have to choose their preferred template on each device, i.e. if the customer selects a template from the web, it will not transfer to iOS.

At the time of the acquisition, Walmart mentioned a social sharing feature that would allow friends to give each other feedback on virtual outfits, but that feature isn’t live. But the retailer says it could be a future virtual trial capability.

During testing, Walmart said it received positive feedback from customers about the experience, which it hopes will make shopping for clothes online more like shopping in person.

“One of the most frustrating parts of shopping for clothing online is figuring out what an item will look like on you. With Zeekit, our goal is to provide an inclusive, immersive and personalized digital experience that will better replicate shopping physical,” said Denise Incandela, executive vice president of apparel and private labels at Walmart in the United States.

Of course, helping customers choose products that fit them well isn’t just about meeting customer needs or encouraging a purchase – it’s about helping to reduce the number of returns on garments that don’t work when they don’t work. are tried at home. . Walmart has attempted to solve the customer hassle part of the problem by offering the ability to return or exchange online purchases at its retail stores. But the ideal scenario would be a total reduction in yields.

Picture credits: walmart

Better testing technology could also help Walmart in its battle with Amazon, which took the lead last year as the top apparel retailer in the United States, ahead of Walmart — a spot it was able to claim, in part, due to the COVID pandemic and the resulting surge in online shopping. Walmart, however, had been hit by “weaker demand” for apparel in the early months of 2021, but restored towards the end of the year thanks to the holiday sales. (It didn’t break out its fourth-quarter apparel sales, though.)

The idea of ​​offering a virtual try-on is something many startups have experimented with over the past decade and beyond, including Fitle, Metail, Virtusize,, and several others. These types of companies have also been subject to acquisitions, such as eBay’s 2014 deal for PhiSix for 3D testing technology, for example. More recently, a startup called was raising money for its computer vision technology that lets customers choose their model to virtually try on clothes.

However, Walmart’s adoption of a virtual try-on feature could push the wider e-commerce industry to incorporate similar technology into its own websites and apps in the future.