On this page
- Technical families
- Platform landscape (as-of 2026-09)
- Google Virtual Try-On
- Snap / Snapchat AR Shopping Lenses
- Returns impact
- Conversion impact
- Market size (as-of 2026)
- Retail adoption examples (as-of 2026)
- In-store extension
- Consumer context
- Web standards layer
- State of Fashion 2026 — McKinsey / BoF
- Key terms
- Next frontiers
Virtual Try-On
Virtual Try-On
Virtual Try-On (VTO) is a technology layer that lets shoppers visualise how a product — apparel, footwear, eyewear, cosmetics, jewellery — will look on their own body or face before completing a purchase. It spans two distinct technical families: generative-AI VTO (project a product image onto a photo using diffusion models) and 3D VTO (render a real-time 3D digital twin on a live camera feed). Both approaches aim to reduce the fit and appearance uncertainty that drives ecommerce returns and pre-purchase hesitation.
Technical families
Haptic Media (2026-06-01) distinguishes two primary architectures:
Generative-AI VTO (image-based diffusion) Projects a product image onto a customer photo or video still using a diffusion model. No 3D product assets required; fast to deploy; produces photorealistic output for most fashion categories. Does not support real-time immersion. Google's TryOnDiffusion model falls into this family — it uses image diffusion to project a garment onto a photo of another model, preserving pose and composition; DressX applies the same technique for fabric rendering. (Haptic Media, 2026-06-01)
3D VTO (real-time AR body mapping) Projects a real-time 3D digital twin onto a live camera feed via a device camera. Immersive and interactive; requires 3D product assets; higher production and maintenance cost. Dominant in eyewear (mature face-tracking algorithms lock onto eyes, nose, lips) and gaining traction in jewellery and watches. Wrist and hand tracking remains significantly harder due to less-defined anatomical anchor points. (Haptic Media, 2026-06-01)
Platform landscape (as-of 2026-09)
Google Virtual Try-On
At Google I/O in May 2025, Google announced a new custom image generation model for fashion VTO that understands how different materials fold, stretch, and drape on different body types; the feature works from a single uploaded personal photo and covers billions of apparel items from the Shopping Graph, which contains over 50 billion product listings (as-of 2025-05-20). (Google Blog, 2025-05-20)
The tool was broadly rolled out on 24 July 2025 across Google Search, Shopping, and Images for billions of items from retailers including Macy's, Kohl's, Walmart, and Nordstrom; the tool stores the shopper's uploaded image (with permission) to avoid re-uploading. Google's Chief Business Officer Philipp Schindler said on the Q2 2025 earnings call: "Early results and engagement have been extremely positive, particularly with Gen Z users." (Digital Commerce 360, 2025-07-24)
According to a September 2024 eMarketer report (as cited by DC360, 2025-07-24), virtual try-on images in Google Search receive 60% more high-quality views than standard product listings, and shoppers try on clothing using an average of four different models per product.
As of 2026, Haptic Media reports Google's VTO API is rolling out across the US and Europe, initially for Diesel and Jil Sander before expanding to Marni and Maison Margiela; it is built into the Gemini Enterprise Agent platform. (Haptic Media, 2026-06-01 — third-party reporting; primary Google source not independently verified)
Snap / Snapchat AR Shopping Lenses
Snap claims that AR shopping lenses can increase conversion rates by 11x compared to traditional shopping methods. (Snap for Business — undated; treat as stale-risk pending a dated primary source)
Amazon Fashion leveraged Snapchat AR Shopping Lenses for eyewear try-on covering thousands of brands including Ray-Ban, Maui Jim, Oakley, and Persol. (Marketing Dive — publication date unconfirmed)
Returns impact
- Perfitly (via Shopify, 2025-08-26) reports customers see return rates decline by up to 64%
- eMarketer (via Bloomreach, 2026-04-20) reports return rates drop by up to 40% for retailers implementing virtual fitting tools
Both figures are from vendor-influenced or vendor-sourced reports. No independent academic or audit study found. Scope (category, deployment quality, control condition) is not disclosed in either case.
VTO is positioned as a direct response to the returns cost problem: Haptic Media (2026-06-01) estimates returns account for roughly $740 billion in global ecommerce annually, though the figure is cited without a named upstream source.
[!unverified] The $740 billion global ecommerce returns figure (Haptic Media, 2026-06-01) lacks an upstream source citation in the article reviewed. Do not treat as primary.
VTO also intersects with Agentic Sizing: without machine-readable fit data, AI agents completing autonomous purchases default to ordering multiple sizes and letting the human sort returns on delivery — scaling Bracketing (Fashion Returns) at machine speed. (YourSizer, 2026-07-29, as cited in the Agentic Sizing concept page)
Conversion impact
- Snapchat AR lenses: 11x conversion (Snap for Business — undated, unverified)
- e.l.f. Cosmetics via Perfect Corp: 200% increase in conversions (DC360, 2025-07-24 — vendor-sourced)
- Estée Lauder via Perfect Corp: 2.5x lift (DC360, 2025-07-24 — vendor-sourced)
- DressX/BoF Professional: 50% higher purchase likelihood among VTO users; luxury conversion up to 10x higher for VTO engagers (BoF Professional search snippet — full report behind paywall; figures unverified)
These figures reflect different product categories (cosmetics vs. luxury vs. apparel), conversion funnel stages (add-to-cart vs. order completion), and control conditions. No independent head-to-head comparison found.
ArtLabs (via Shopify, 2025-08-26) states brands using AR experiences are 41% more likely to capture consumer attention compared to those that don't.
Market size (as-of 2026)
- $4 billion in 2024 → $32.7 billion by 2034 (22.9% CAGR) — Global Insight Services, via Haptic Media 2026-06-01
- $5.8 billion in 2024 → $27.7 billion by 2031 — Valuates, via Digital Commerce 360, 2025-07-24
- $12.09 billion in 2025 → $38.92 billion by 2030 (26.3% CAGR) — The Business Research Company, via Bloomreach, 2026-04-20
Scope differences (virtual try-on only vs. broader virtual shopping; fashion-only vs. all categories; online vs. all channels) are not consistently disclosed. The 3x variance between the lowest and highest 2024/2025 baseline figures reflects definitional disagreement more than measurement error.
Retail adoption examples (as-of 2026)
| Brand / Retailer | Category | VTO type / notes |
|---|---|---|
| Adidas, ASOS, Macy's | Apparel | Virtual fitting room (Shopify, 2025-08-26) |
| Zalando, Net-a-Porter | Apparel | ML / computer vision virtual fitting (Shopify, 2025-08-26) |
| Chanel | Makeup, eyewear, jewelry, watches | Full category VTO on website (Haptic Media, 2026-06-01) |
| Cartier | Watches, jewelry | VTO directly on website (Haptic Media, 2026-06-01) |
| Fendi, Burberry, Prada, Gucci | Luxury apparel / accessories | AR technologies implemented (Shopify, 2025-08-26) |
| Diesel, Jil Sander | Apparel | Google VTO API pilot — EU first wave (Haptic Media, 2026-06-01) |
| Amazon Fashion | Eyewear | Snapchat AR Shopping Lenses integration (Marketing Dive — date unconfirmed) |
In-store extension
Haptic Media (2026-06-01) reports VTO is increasingly integrated into physical stores via connected mirrors, allowing customers to try products or sizes not physically in stock — including colour variants, custom engravings, and made-to-order items — and to order on the spot. This extends VTO from a returns-reduction tool to an in-store inventory-expansion mechanism.
Consumer context
EY Future Consumer Index (2024), as cited by Haptic Media (2026-06-01), found 57% of consumers want to see, touch, and feel items before buying.
Retail Dive (undated, via Haptic Media 2026-06-01) found 92% of European shoppers favour physical stores specifically to see, touch and try products — framing VTO as a bridge to digital confidence rather than a replacement for physical experience.
University of Illinois researchers are developing VTO technology framed as reducing "online shopping stress" — positioning fit uncertainty as a shopper anxiety problem rather than purely a conversion metric. (YouTube/local media segment, 2026-04-17)
Web standards layer
The WebXR Device API has reached W3C Candidate Recommendation Draft status, enabling browser-native AR and VR without app installs — a foundational capability for browser-based virtual fitting rooms. Chrome's HTML-in-Canvas API (origin trial) allows interactable web UI to be placed into 3D scenes within immersive WebXR experiences. (web.dev / Chrome Developers Blog search results, ~2026; exact article dates unconfirmed)
State of Fashion 2026 — McKinsey / BoF
The State of Fashion 2026 report (McKinsey & Company / Business of Fashion, published 2025-11-17, 143 pages, PDF behind registration gate) contains a dedicated "AI Shopper" chapter covering AI-powered virtual try-on as a core consumer shift; it discusses generative AI enabling shoppers to use VTO before purchase and AI agents acting as style consultants. See _research/report-queue.md for signup queue entry.
Key terms
| Term | Meaning |
|---|---|
| Generative-AI VTO | Image-based diffusion model that projects a product onto a shopper photo; no 3D assets required |
| 3D VTO | Real-time AR body mapping via live camera; requires 3D product assets |
| TryOnDiffusion | Google's specific diffusion model for garment projection onto model photos |
| WebXR | W3C browser API enabling AR/VR without app installs |
| Agentic Sizing | Machine-readable fit data enabling AI agents to resolve sizing autonomously |
Next frontiers
- WebXR — W3C browser standard for AR; VTO delivery mechanism
- TryOnDiffusion — Google's underlying model technology
- True Fit — fit intelligence vendor with MCP API for agentic commerce
- Perfect Corp — VTO vendor behind several cited conversion studies
- DressX — digital fashion VTO platform; BoF Professional conversion report