DMEXCO 2026 sets Scaling Intelligence theme for AI in production

DMEXCO 2026 Scaling Intelligence is the theme for the next edition of the Cologne event, scheduled for September 23 and 24, as organizers shift the program toward AI systems that are operating in production. The agenda is being framed around cases where AI is tied to revenue, retention, margin or operating efficiency, with fashion e-commerce virtual try-on presented as one example of a category moving past pilots.

Virtual try-on has been discussed in retail for years, but earlier versions depended on costly 3D product work and often struggled with ordinary customer images. The newer claim from vendors and retailers is that generative AI can create usable 3D representations from standard product photography and render fit, fabric and movement more quickly than older workflows allowed.

Several deployments are already live. ASOS lets shoppers upload a photo or create a digital twin based on their measurements and stated preferences. Breuninger has integrated Google’s virtual try-on technology into its app, becoming the first German fashion retailer to do so, according to the companies. Maybelline offers shade testing through uploaded images, a digital model or a live camera view.

What is DMEXCO 2026 Scaling Intelligence?

DMEXCO’s “Scaling Intelligence” theme refers to AI use cases that have moved from experimental demos into day-to-day business systems. For operators and investors, the test is whether a model performs for real users, under real constraints, and produces numbers the business already tracks.

The retail case is useful because the economics are concrete. Online fashion has long faced conversion friction from fit uncertainty and high return volumes. The National Retail Federation estimated that 19.3% of online fashion purchases in the US were returned in 2025, while Gen Z shoppers returned close to eight garments on average. Any tool that reduces uncertainty before checkout has a plausible path to margin impact, though individual retailers have not disclosed comparable revenue or return-rate results in the examples cited.

Arnold Pötsch, lead author of the BVDW working group paper on 3D in e-commerce, said advances in computer vision, AI and real-time rendering have made virtual try-on a competitive advantage for retailers that adopt it. He also linked the technology to digital twins, more personalized shopping, lower returns and sustainability gains.

DMEXCO is extending that pattern beyond commerce. Its 2026 program will cover AI adoption across fintech, healthtech, deeptech, SaaS and sustainability, including the event’s World of Commerce, World of Tech and dedicated summits. The premise is that underwriting, fraud detection, diagnostics, triage, robotics, materials work and software retention are being judged less by feature announcements and more by whether deployments hold up in production.

The event format is also positioned around comparison. DMEXCO says the Expo floor will put scaled products from established companies near early-stage companies in the Start-up Area, while conference sessions will focus on what changes when teams move models from pilots into live systems. For European companies, the discussion carries a regional angle: many builders are trying to scale AI under different regulatory and capital conditions than US platforms or Chinese manufacturers.

The unanswered piece is measurement. Virtual try-on has clearer links to returns and conversion than many AI claims, but companies still need to disclose baselines, deployment costs and sustained outcomes for the market to separate infrastructure from marketing. DMEXCO 2026 is betting that those comparisons will define the next phase of AI adoption.

This story draws on original reporting from Tech.eu.