Google’s new AI try-on: what changed
Google this week expanded its AI-powered apparel try-on capabilities so users can preview clothes with just a single selfie. The feature, integrated into Google Search and Google Shopping, leverages the company’s vision and generative models to map garments onto a user’s photo — removing the need for full-body scans or multiple photos. Google says the update aims to speed up online discovery and cut into the high return rates that plague e-commerce fashion.
Background and how it works
Virtual try-on is not new: retailers and platforms have long used AR and body-scanning tech to let shoppers visualize products. Snapchat, Meta and platforms like Shopify have offered augmented-reality try-on tools for cosmetics, eyewear and limited apparel categories. Google’s approach combines image understanding with cloth rendering to infer body pose and approximate measurements from a single selfie, then simulates how a garment would sit and drape on that silhouette.
Google ties the feature into existing product listings in Search and Shopping results, so shoppers can tap a “Try on” button on eligible items. The experience runs on both mobile web and native apps and, in many cases, does the computation on-device or via Google’s cloud-based models to generate a realistic preview. Google also points to its investment in the Shopping Graph — its product database that aggregates retailer inventory, sizes and images — as a backbone for matching real SKUs to try-on previews.
Data points and industry context
Online fashion returns historically sit between 20% and 40% depending on category; misfit and unmet expectations are leading causes. Retailers including ASOS, Zara-owner Inditex and H&M have experimented with fit and size tools to reduce returns and improve conversion. Analysts estimate AR and AI try-on can improve conversion rates by low double digits while reducing return costs — a major incentive given the logistics and sustainability implications of high return volumes.
Implications for shoppers and retailers
For consumers, the single-selfie model lowers the barrier to trying clothes virtually. It removes friction compared with body-scans or multi-angle uploads and could boost adoption of AI try-on in mainstream shopping. For retailers, tighter integration with Google Search and Shopping means broader exposure but also greater pressure to supply high-quality images, accurate size metadata and consistent SKUs in the Shopping Graph.
There are trade-offs. Single-image inference can’t match the fidelity of a full 3D scan; fit precision — especially around shoulders, hips and sleeve length — will still vary. Retailers relying on precise size recommendations may need complementary measurement data or fit profiles. Privacy is another concern: while Google says some processing occurs on-device, users and privacy advocates will scrutinize how selfies and derived body metrics are stored and used.
Expert perspectives
“Lowering the input requirement to a single selfie removes a major UX hurdle,” said a retail technology analyst familiar with AR commerce, noting that convenience is a primary adoption driver. “However, brands will still need robust size data and clear return policies to build long-term trust.”
Another industry observer, who consults for fashion retailers, added: “Google’s distribution advantage — Search and Shopping — is the real lever here. If more shoppers can test an item before they click through, conversion should improve. But accuracy limitations mean this is best seen as a discovery tool, not a replacement for traditional sizing charts and fit tools.”
Risks, regulation and competitive landscape
Regulation and standards around digital body data are emerging. Lawmakers in Europe and parts of the U.S. have raised questions about biometric data and automated profiling. Companies will need to be transparent about retention, consent and how inferred body metrics might be used for advertising or personalization.
Competition is fierce. Snap has long promoted AR try-on for fashion and beauty; Meta is pushing its own commerce integrations; and specialist vendors like Fits Me (now part of other platforms) and 3D modeling startups provide deep-fit solutions for enterprise retailers. Google’s advantage is scale — the company can route millions of search queries through a try-on layer if retailers opt in.
Conclusion: what’s next
Google’s selfie-driven try-on lowers the bar for mainstream adoption of virtual apparel previews and could nudge shoppers toward richer, AI-assisted product discovery. The technology is likely to accelerate investment in size metadata, high-quality product imaging and privacy safeguards as retailers and regulators adapt. For now, expect retailers to test the feature’s conversion and return effects, while competitors and startups race to offer more precise 3D fit and integration options.
Related topics: Google Shopping updates, Google Lens, AR commerce, generative vision models, e-commerce returns reduction.
About the Author
