At Innovation Day UK 2026, Roland Butler, Head of Product within Zalando’s Content Solutions business, offered a behind-the-scenes look at one of Europe’s most advanced content operations.
Zalando is often described through its logistics scale or assortment size. Roland’s keynote focused on content as infrastructure and the practical realities of using AI to deliver inspiration to millions of customers every day.
From product pages to inspiration engines
Roland opened with a simple observation. Product Detail Pages started as functional records: one product, a white background, a price. That model no longer holds.
Many customers now arrive looking for ideas. They expect inspiration and context. At Zalando’s scale, that creates a structural challenge. How do you offer richer experiences without multiplying cost, production time and organizational complexity?
That question sits at the heart of Content Solutions, the business unit responsible for all content across the Zalando platform. Its mission is to deliver new experiences with AI on a massive scale, with dramatically lower marginal costs. At the same time, it aims to drive significant topline impact through improved inspiration and richer information, personalized to the customer, and leading to the right purchase decisions.
Proving formats before scaling them
Zalando’s AI operating model is deliberately demanding. New formats are explored in an innovation lab, where teams can test ideas and workflows in controlled conditions. Only when a format has proven its value does it move into the core system.
At scale, a good idea is meaningless if it can’t be executed repeatedly without failure. Roland was openly skeptical of vendors who promise transformation without handling real platform conditions (scale, data complexity, etc). At Zalando, that means running correctly thousands of times a day, on a live platform with very high customer expectations. Also, in a space that is evolving this fast, what you will find yourself doing is transferring your domain knowledge to the vendor and making their organisation smarter rather than your organisation..
One early example is AI-generated backgrounds that allow imagery to adapt to different markets, audiences or test variants.
Curated stories instead of static pages
Beyond PDPs, Zalando is reshaping its home feed into a curated, personalised, and narrative-driven surface. Less catalogue, more editorial logic.
These curated stories increasingly use AI-generated content and are assembled dynamically based on audience, behavior and context.
“Customers don’t care how an image was produced. They care whether it inspires them and helps them to make a choice.”
Roland Butler, Zalando
To enable this, Zalando has invested heavily in proprietary tooling with a multi-model approach. Different engines are used depending on quality requirements, cost constraints and use cases. Models can be swapped, combined or deprioritized without blocking delivery.
Creative teams at a different altitude
AI has not removed creative judgment but amplified its consequences. Instead of crafting individual assets, creative leads define systems, styles and rules that AI executes at scale. With the same team size and in a single week, Zalando can now shape the look of product lines on a massive scale. This speed allows teams to react to microtrends in under 24 hours.
“AI didn’t make our creative teams less important, it made their decisions matter more, because one choice now shapes thousands of products.”
Roland Butler, Zalando
The same principle applies to video. As a platform business, Zalando receives hundreds of thousands of videos from partner brands. AI analyzes inbound content, checks alignment between imagery and product attributes, and routes only edge cases for human review.
Attribute enrichment as a growth lever
Attribute enrichment is foundational at Zalando’s scale. Even basic attributes such as color or garment type can’t be handled manually with sufficient consistency.
AI automates this process end to end, including localization. In 2025, Zalando achieved 1.8 million AI-enriched PDP attributes. It enables deep tagging of every associated media asset. Every image, every video, every render becomes searchable and targetable.
This creates a media library where performance marketing teams can activate with precision. The commercial impact is measurable. Improved relevance drives higher engagement, increases conversion and contributes directly to GMV uplift. It’s a level of enrichment that would be impossible with a purely human model.
Reimaging the production pipeline
At scale the small things matter: a human QA loop for hundreds of product images is no issue, nor is manual retouching. But when you are faced with thousands of images per day, it’s a different story.
Zalando’s acquisition of a dedicated 3D company ShopAR helps it to rethink production from first principles. Products such as shoes are scanned once at high fidelity, then activated across a single 3D pipeline. AI-generated environments are layered around these scans, enabling rapid variation without repeating the entire shoot process. Repetitive work is automated, while human attention shifts to creative direction and quality control.
The same principles apply to campaign content and PDP production. The impact on throughput is significant. Production time for scanned items has dropped from five days to under two, while costs are down and quality has improved. The same goes for campaign content. What used to take weeks can now be done in 10 minutes.
Three lessons that travel beyond Zalando
Roland’s keynote was rich in examples and metrics. But three lessons stand out for leaders trying to scale content without scaling chaos.
1. Don’t allow a divide between technology and creative team
Zalando’s results are driven by better tools but also tight alignment. Creative intent is defined upstream and carried through the system. Technology doesn’t replace creative judgment. It operationalizes it. Orchestration only works when both sides design the system together.
2. Resist the comfort of single-vendor certainty
Zalando’s multi-model strategy avoids dependency on a single provider. Teams can act as the market shifts, while decisions and production keep moving. This avoids a common failure: waiting for the ‘right’ model, platform or standard to emerge. By the time it does, the organization has already fallen behind.
3. Be brave and be fast
This space is evolving incredibly fast. Customers' expectations are changing, the regulatory environment is evolving, cost models are changing. For sure you can wait until it stabilizes, but your competitor won't.
Taken together, these lessons point to a broader shift. AI is no longer a capability you add at the edges. It is something you design into the way content is produced, governed and scaled. Zalando is not experimenting with AI. It’s running on it.
Contributor in this article
