It pinches at the shoulder. it drags at the hem; the draping simply doesn’t fall the way the professional photography suggested. For millions of online shoppers, this sequence of disappointments is a weekly ritual. For the retailers who ship those garments, it is a financial hemorrhage.
The fashion industry is currently locked in a battle with a phenomenon many executives call its “silent killer”: the escalating cost of product returns. To stem the tide, a new wave of companies is deploying AI to solve retail’s returns problem, moving beyond simple size charts toward “digital twins” and physics-based simulations that attempt to replicate the tactile reality of a dressing room.
The scale of the crisis is staggering. According to estimates from the National Retail Federation (NRF), consumers were expected to return nearly $849.9 billion in merchandise in 2025, representing roughly 15.8% of all annual retail sales. The problem is significantly more acute in the digital space, where the return rate for online sales jumps to 19.3%.
Much of this volatility is driven by Gen Z. Shoppers between the ages of 18 and 30 averaged nearly eight online returns per person last year, according to NRF data. This behavior is often fueled by “bracketing”—the practice of ordering the same item in multiple sizes or colors with the intent of returning most of them.
Moment Makers Group | Istock | Getty Images
The high cost of “free” returns
For years, free returns were viewed as a competitive necessity to lure customers away from physical stores. However, that convenience has become a structural liability. In many cases, the cost of processing a return—shipping, inspection, and restocking—exceeds the actual value of the refund. Many returned items never even produce it back to the shelves, ending up in landfills or liquidators.
This margin erosion is happening against a backdrop of broader economic instability. Retailers are currently navigating a complex environment where U.S. Trade policies and tariffs have increased the cost of goods manufactured in Southeast Asia. When combined with inflationary pressures that make consumers more price-sensitive, the cost of a 19% return rate becomes unsustainable.
Retailers are now forced to balance a difficult contradiction: 82% of consumers consider free returns essential to their purchasing decision, yet the financial burden of providing them is eating directly into gross margins.
From “pretty” pictures to digital physics
While virtual try-on technology has existed since the 2010s, early iterations were often little more than digital stickers—images of clothes overlaid on a photo of a person. The emergence of generative AI and increased computing power has shifted the goalpost from visual approximation to “mirror-like realism.”
One example is the startup Catches, which allows users to create a “digital twin” for virtual fittings. Unlike previous models that Ed Voyce, founder and CEO of Catches, says “just look pretty,” this platform incorporates the actual physics of fabric texture and how different materials interact with a moving human body.
The technical leap is made possible by a confluence of high-end infrastructure. Catches is built on Nvidia’s CUDA platform and backed by Antoine Arnault of LVMH. According to Voyce, the technology is finally viable because brands can now run these complex visuals in the cloud on “bare metal” cheaply enough to notice a meaningful return on investment.
A bifurcated strategy: Tech vs. Policy
Retailers are currently split between two primary methods of protecting their margins: implementing stricter return policies or investing in AI-driven accuracy.
| Approach | Primary Method | Example Retailer | Goal |
|---|---|---|---|
| Policy-Led | Return shipping fees | Zara | Discourage “bracketing” |
| Tech-Led | Virtual try-on (VTO) | ASOS / Shopify | Reduce sizing uncertainty |
| Hybrid | Fees + VTO tools | Zara | Protect gross margins |
Zara, owned by Inditex, was an early adopter of return fees for online orders. While the move was contentious among customers, it served as a direct deterrent to the “buy three, return two” habit. Simultaneously, the retailer launched “Zara try-on” in December to bridge the gap between the screen and the mirror.
Other players are leaning harder into deep-tech partnerships. ASOS has experimented with virtual try-ons via the startup AIUTA, allowing customers to visualize clothing on a diverse array of body types, heights, and skin tones. The results have been tangible; ASOS recently noted an improvement in profitability partially driven by a 160 basis point reduction in its returns rate.
The democratization of this technology is also being driven by platform giants. Shopify has integrated Genlook’s AI virtual try-on app into its commerce ecosystem to boost buyer confidence, while Google has integrated virtual try-on capabilities directly into product search results as of April 30.
The limits of the algorithm
Despite the optimism surrounding digital twins and generative AI, industry analysts warn that technology cannot fix a fundamental product failure. Simeon Siegel, Senior Managing Director at Guggenheim, notes that while these tools bridge the gap, they will never fully replace the physical experience of trying on a garment.
“What you sell is always going to be more important than how you sell, and so I just suppose remembering that will help dictate who wins and benefits and amplifies from AI versus who gets consumed by it,” Siegel said.
Beyond the fitting room, retailers are expanding their use of AI into inventory management and fraud prevention to further protect the bottom line. The goal is a holistic reduction in waste, rather than a reliance on a single “magic wand” solution.
As the industry moves forward, the next critical checkpoint will be the 2025 year-end retail data, which will reveal whether the integration of virtual try-ons across Google and Shopify platforms has meaningfully lowered the global return percentage or if consumer habits—particularly among Gen Z—remain resistant to digital intervention.
This represents a developing story. We invite our readers to share their experiences with virtual try-on tools in the comments below.
Disclaimer: This article contains financial analysis and market data for informational purposes only and does not constitute investment advice.
