AI Virtual Try-On Is Taking On Retail's $849 Billion Returns Problem — Google Embeds It Directly in Search Results April 30

The US National Retail Federation estimated that 15.8% of annual retail sales were returned in 2025, totaling $849.9 billion. For online purchases specifically, the return rate jumped to 19.3%. Gen Z shoppers aged 18–30 averaged nearly eight online returns per person. Most returned items never make it back to shelves, and processing a return often costs the retailer more than the refund itself. AI virtual try-on is now the industry's leading bet to reduce this problem where it starts: the moment a shopper looks at a product photo and thinks, "but will it fit me?"

Who Is Launching What

Google will make virtual try-on technology accessible directly within product search results across its platforms starting April 30. Users upload a full-length photo and receive an AI-generated image of themselves wearing the item within seconds. The feature works with billions of product listings from Macy's, Kohl's, Walmart, and Nordstrom through Google's Shopping Graph, which indexes more than 50 billion products. Google also launched the Doppl app, which converts static try-on images into short video clips so shoppers can see how an outfit looks in motion.

Other major retailers have already moved. Zara launched its "Zara try-on" tool in December 2025. ASOS partnered with deep-tech startup AIUTA to let customers see clothing on a range of body types, heights, and skin tones. ASOS reported a 160 basis point improvement in its returns rate, citing virtual try-on as a contributing factor. Nike Fit uses computer vision to scan feet and recommend the correct shoe size for each specific model. Shopify integrated Genlook's AI virtual try-on app across its commerce platform, making the technology available to independent merchants.

The Startups Building the Fit Layer

Behind the big retail names, a wave of startups is building the underlying technology. Catches, backed by LVMH's Antoine Arnault, creates a "digital twin" that simulates the physics of fabric texture and how material interacts with a moving body. Built on Nvidia's CUDA platform, Catches is live on luxury brand Amiri's website and projects a 10% increase in conversions and 20–30x ROI for brand partners.

AIUTA, the deep-tech partner powering ASOS's implementation, focuses specifically on diverse body representation — ensuring try-on results are accurate across different body shapes rather than defaulting to a narrow set of model physiques.

Early data from eMarketer suggests the technology changes shopping behavior. Virtual try-on images in Google Search receive 60% more high-quality views than standard product listings. Shoppers try on clothing using four different models per product on average and are more likely to visit the brand's website after engaging with try-on results.

The global virtual try-on market reached $5.8 billion in 2024 and is projected to hit $27.7 billion by 2031, according to Valuates Reports.

Can the Technology Actually Deliver on Accuracy?

The technology is improving but not yet at a point where retailers are willing to guarantee fit. ASOS cautions that its tool provides "general guidance" and that customers should still check size guides before purchasing. The primary reason for both returns and abandoned carts remains uncertainty over fit, according to Catches CEO Ed Voyce.

Retail analyst Sucharita Kodali at Forrester notes that the benefits of virtual try-on are directionally clear, but quantifying the exact return reduction remains difficult for many retailers. Attribution is messy — a shopper who uses try-on and keeps an item might have kept it anyway.

What is not in dispute is the scale of the problem these tools are targeting. At $849.9 billion in annual returns, even a single-digit percentage reduction represents tens of billions in recovered revenue.

Sources: CNBC (April 5, 2026); National Retail Federation (2025 data); eMarketer (September 2024 report); Valuates Reports (market sizing); Google Shopping blog.