Building the Visual AI Layer for E-commerce

Our talk at the Nanjing Trade Fair (金洽会) was an introduction to Youzu for an audience of Chinese retailers, partners, and investors — and an honest tour of where we think the visual AI layer for e-commerce is going.

The opening

"We're a visual AI platform helping e-commerces with product discovery and immersive experiences. 73% of shoppers want better tools. Conversion rates are low. Cart abandonment is a problem. We came up with in-house solutions designed specifically for e-commerce."

The lens reframe

The talk repeats one of our favorite framings:

"You're probably familiar with Google Lens, Temu Lens, AliExpress Lens. But 99% of e-commerces don't have access to these tools. This is the future. Gen Z already shops this way — they like visual experiences. And don't think of Lens just as product discovery. Think of it as a recommendation engine. So far, recommendations have happened through text. We're enriching that data with visual signals."

What makes Room Visualizer different

"We're the first startup to give the whole end-to-end customer experience. Most platforms just generate an image. We generate the image — and then we let the customer click on the table inside that image and see similar items they can actually buy from the catalog."

The market

"Just considering interior design, furniture e-commerce, and visual search — we're already approaching a $3 trillion market. If you take the whole e-commerce industry, it's much more."

On differentiation

The honest version of the competitive story:

"Our competitors do one single point feature. We're way ahead because our infrastructure costs are way lower — that's our innovation. They cost more, and they don't give the whole platform."

Onboarding velocity

"Onboarding is roughly four days. But it depends on the catalog. Billions of SKUs — yes, four days. Smaller catalogs, one day. In the smallest cases, a few hours. We provide both API and SDK."

The team

Berlin-based. Chief AI Officer Omar Abdul as CTO, Florian as founding engineer, and Nail at the helm — all veterans of Sportotal, where the team built AI broadcasting cameras for over 1,000 sports clubs. The same computer vision DNA, applied to a much bigger problem.

"We're not limited by geography. We can work with any customer globally — and we'd love to partner with Chinese companies."