What the AI Blunders at Bol, Omoda and Zara Really Tell Us

This weekend, De Telegraaf published an article highlighting a series of painful AI mistakes at well-known retailers, including Bol, Omoda and Zara. Among the examples were product titles that accidentally included ChatGPT’s own response, AI-generated fashion models with obvious physical distortions, and a Zara product where the zodiac sign Cancer was translated into Dutch as the disease “kanker” instead of the constellation Kreeft.

While these examples are amusing at first glance, they reveal a much bigger challenge facing retailers as AI becomes part of everyday ecommerce.

You can read the original article, “Webwinkels als Bol vol met pijnlijke AI-blunders: ‘Dit is gewoon gênant’”, here.

Looking Beyond the Headlines

The obvious reaction is to blame AI. After all, AI generated the image, wrote the product title, or translated the product description. But these incidents aren’t really AI failures.

They’re process failures.

Artificial intelligence doesn’t decide what should be published. It doesn’t know whether a translation aligns with your brand terminology, whether a generated image still accurately represents the product, or whether a product title accidentally includes an internal ChatGPT response. AI simply generates content based on the information it receives and the instructions it is given.

In other words, AI is doing exactly what it was designed to do.

AI Only Knows What You Tell It

Retailers are increasingly using AI to write product descriptions, generate titles, translate content, enrich product data, and even create product imagery. The productivity gains are enormous, allowing businesses to launch products faster and create content at a scale that would have been impossible just a few years ago.

However, AI has one important limitation: it cannot determine whether the underlying product information is complete, accurate or trustworthy. If the source data contains inconsistencies, missing attributes or incorrect terminology, AI will simply build upon those imperfections. Instead of correcting mistakes, it scales them.

That’s exactly what makes these examples so interesting. They are not isolated AI mistakes; they are visible symptoms of a much deeper issue.

Product Data Has Become Business Critical

As AI becomes embedded in every stage of the product content lifecycle, the quality of product data becomes more important than ever.

Customers don’t distinguish between a mistake made by AI and one made by a person. They simply see incorrect product information, confusing translations or unrealistic images. Every error reduces confidence in the brand and makes it easier for customers to buy elsewhere.

The De Telegraaf article demonstrates just how quickly these issues become public. What might once have been a minor typo can now spread across webshops, marketplaces, marketing campaigns and AI shopping assistants within minutes.

The faster AI works, the more important it becomes to ensure that the information it uses is reliable from the start.

The Missing Foundation

The discussion shouldn’t be about whether companies should use AI. AI is here to stay, and it will continue to transform ecommerce.

The real question is whether organizations have built the right foundation to use it responsibly.

That foundation is validated and verified product data.

Validated product data ensures that information is complete, consistent and complies with predefined business rules before it reaches customers. Verified product data goes one step further by confirming that the information accurately represents the real product. Together, these processes provide the trusted data that AI needs to generate high-quality content and deliver reliable customer experiences.

Without that foundation, AI simply automates mistakes.

With it, AI becomes a powerful accelerator for productivity, consistency and customer satisfaction.

The Real Lesson

The AI blunders highlighted by De Telegraaf are not a warning against using artificial intelligence. They are a reminder that AI is only as good as the data behind it.

Organizations that focus solely on implementing the latest AI tools risk scaling poor-quality product information faster than ever before. Those that invest in validated and verified product data will be able to unlock AI’s full potential while maintaining customer trust.

In the end, successful AI doesn’t start with better prompts.

It starts with better product data.