Google Is Shutting Down Doppl on April 30 — But AI Virtual Try-On Is Moving Directly Into Search

Doppl closes April 30. Virtual try-on moves into Google Search on the same day. CNBC reports AI startups are targeting retail's $700B+ returns problem.

April 5, 2026 · 7 min read · By Connie

What Is Doppl and Why Is Google Shutting It Down?

Google Doppl is a Google Labs experiment launched in June 2025. The app allowed users to upload a photo of themselves and virtually try on clothing from major retail partners — tops, bottoms, and dresses — with AI-generated previews showing how garments would look on their specific body. It was designed to address a fundamental problem in online fashion retail: customers cannot try clothes on before buying, leading to return rates of 20-40% for apparel.

Google is shutting down Doppl on April 30, 2026, not because the technology failed — but because the experiment succeeded. The virtual try-on capability proven in Doppl is being integrated directly into Google Search and Google Shopping. From April 30, users can access virtual try-on by tapping on product listings or image results and clicking "Try It On," without needing a standalone app. The technology becomes a feature of the world's largest search engine rather than a niche experiment.

This is a significant distribution upgrade. Doppl required a separate download and active user intent to use. Google Search integration means virtual try-on is now one click away for any Google Shopping user — potentially hundreds of millions of people — during the normal shopping discovery process.

The Returns Problem: Why Retailers Are Investing in AI Try-On

Retail returns are one of the most expensive problems in commerce. CNBC reported today on AI startups specifically targeting what the industry calls "silent killers" — the hidden cost of processing, restocking, and writing off returned merchandise.

The scale of the problem is staggering. US retail returns are estimated to cost the industry over $700 billion annually in direct costs and lost revenue. Online fashion returns are especially severe, with average return rates of 20-40% for apparel — compared to under 10% for in-store purchases. The difference is largely driven by uncertainty about fit and appearance that customers cannot resolve without trying clothes on.

Every returned item costs retailers on multiple dimensions: reverse logistics, processing labor, quality inspection, repackaging, and often a markdown on resale value. High-volume returns can make profitable-looking product lines actually lose money when total costs are accounted for. This is the problem AI virtual try-on is designed to solve.

Who Is Using AI Virtual Try-On in 2026

The Macy's data point is instructive. Macy's launched its AI stylist chatbot "Ask Macy's" on March 23, 2026, powered by Google Gemini. Customers who engage with the AI assistant spend 4.75x more per visit than those who do not. The combination of conversational AI and visual try-on technology is producing measurable commercial outcomes — not just convenience improvements.

The AI Startups Solving Returns at Scale

Beyond Google's infrastructure play, CNBC's reporting today highlights a wave of smaller AI startups attacking the returns problem from different angles. These "silent killer" solvers are building technology that helps retailers reduce return rates before merchandise ever ships.

Fit prediction models analyze historical purchase and return data combined with user body measurements to predict which size a customer should order, reducing sizing-driven returns. Fabric simulation technology shows how different materials drape, move, and wrinkle on different body types. Occasion matching AI helps customers understand whether a garment is appropriate for their stated use case, reducing "not what I expected" returns.

The addressable market is enormous. Even a 5 percentage-point reduction in online fashion return rates — from 30% to 25% — would save the industry tens of billions of dollars annually. Retailers are now treating AI try-on and fit prediction as infrastructure investment, not a nice-to-have feature.

What Doppl's Shutdown Tells Us About Google's Strategy

Google Labs experiments follow a consistent pattern: launch as a standalone experiment, measure user behavior and technical performance, then either shut down (if it fails) or integrate into core products (if it succeeds). Doppl's graduation into Google Search is a strong signal that virtual try-on performed well enough to justify the infrastructure cost of embedding it across all of Google's shopping surfaces.

This matters for retailers. Google Shopping is the primary price comparison and product discovery tool for a large share of online retail traffic. Having virtual try-on natively embedded in Google Shopping results means the technology will appear in the shopping journey for customers who would never have downloaded a standalone try-on app. The reach is qualitatively different.

For retailers not yet integrated with Google Shopping's try-on program, the urgency is increasing. Competing product listings without try-on capability will appear less engaging next to listings that offer an AI visualization. As try-on becomes table stakes on Google Shopping, retailers without it face a disadvantage in conversion rate.

Frequently Asked Questions