Visual AI in Retail: Three Strategies for Personalization

With its AI-powered discovery feed in the Zalando app, Zalando brings personalized product recommendations, inspiration, and entertainment formats together in one central place. By October 2025, the feed was already live in 22 countries. Instead of a traditional home screen, users see a personalized feed that combines curated boards, individual product recommendations, shoppable livestreams, and editorial content from brands, creators, and Zalando itself. The feed is complemented by new public customer profiles: a personal space for saved content, followed brands and creators, and users’ own boards, which can be shared with others.

With this approach, Zalando wants to evolve its app from a pure e-commerce platform into a content and community platform that combines personalized product recommendations with inspiration, entertainment, and social networking — essentially creating a social commerce app. Co-CEO Robert Gentz sums up the strategy by saying that Zalando wants to evolve from a transactional platform into a “dynamic, inspiring ecosystem for fashion and lifestyle.”

Debenhams Group uses Amazon Web Services (AWS), automated product data, and its Interactive AI Room Styler to prepare product information at scale and enable visual personalization. According to AWS’s announcement about its collaboration with Debenhams Group, the online retail group behind the Debenhams, PrettyLittleThing, boohoo, boohooMAN, and Karen Millen brands expanded its partnership with AWS through a multi-year agreement.

Using Amazon Bedrock, Debenhams automates product descriptions and translations for tens of thousands of items — according to the company, twenty times faster than before, including automatic translation into six languages. These detailed product data provide an important foundation for making products quickly available in new markets, discovery experiences, and personalized product recommendations.

This automated product data also forms the basis for applications such as the Interactive AI Room Styler, developed with AWS: Customers upload a photo, select an interior design style, and receive AI-powered decor recommendations that are directly linked to purchasable products.

CEO Dan Finley describes the partnership as part of a long-term strategy centered on a scalable, cloud-first architecture.

Clienteling refers to personalized, technology-enabled customer service across digital and physical touchpoints. In May 2026, OTB Group and Google Cloud announced an AI-powered virtual try-on solution designed as a clienteling tool.

The fashion group behind Diesel, Jil Sander, Maison Margiela, Marni, and Viktor&Rolf initially planned to roll out the solution for Diesel and Jil Sander in the U.S. and Europe, before expanding it to Marni and Maison Margiela.

The application is designed as a tool for sales teams and is intended to connect digital and physical customer journeys. Sales associates can show customers hyperrealistic, personalized visualizations of individual items and use them as part of one-to-one consultations.

The underlying virtual try-on API runs on Google’s Gemini Enterprise Agent Platform and is complemented by image editing and video generation capabilities. Nano Banana, Google’s image model from the Gemini family, can selectively edit photos and place consumers into campaign imagery. This is followed by Veo, Google’s AI-powered video generator, which can turn static images into short, high-quality video clips.

OTB founder Renzo Rosso says the concept reflects a vision he has pursued for three years: AI should strengthen human customer service rather than replace it. Google Cloud President Matt Renner makes a similar point, saying the goal is to equip sales teams with tools for “deeply personal, confident shopping experiences.”

The example shows that digital AI applications and personal customer service do not have to be opposites. Visual tools can help sales teams provide more individualized service across different customer touchpoints.

What all three examples demonstrate is that product decisions, personalization, and customer service are increasingly being enhanced through visual formats. A feed of curated content boards, a virtual room filled with matching decor, or a garment shown through virtual try-on: Visualizations can illustrate how products might fit into different situations, styles, or usage contexts.

Product descriptions, structured data, and product attributes remain an important foundation. Visual AI adds another layer to this information: It places products in concrete contexts and can turn a traditional recommendation into a “This is what it could look like” experience.

Companies that want to make products more tangible and engaging need more than AI — they also need people. None of the three examples described here focuses on completely replacing human employees. Zalando continues to rely on creators and its community as sources of inspiration, Debenhams automates repetitive tasks in the background, and OTB specifically equips sales teams with new tools to support customer consultations.

The examples therefore follow a similar pattern: AI primarily handles scaling, automation, and visualization, while people continue to play an important role in customer service, curation, and relationship-building.

For retailers and brands, one thing is becoming clear: The future of personalization will not be shaped by algorithms and databases alone, but increasingly by visual experiences as well. Where visual inspiration meets human expertise and advice, additional value can be created for the shopping experience.