Because AI will not only look at the story on your homepage. It needs structured information it can compare, classify, validate, and explain. If the data behind your products, services, categories, and systems is messy, the nicest brand story in the world may not be enough.
That is where data becomes a visibility problem, not just an internal operations problem. Yesterday, SEO helped companies get found. Tomorrow, structured data may decide who gets chosen.
I recently heard someone describe the shift as: yesterday you needed SEO, today you need GEO. GEO stands for Generative Engine Optimization. Maybe that term stays. Maybe someone invents a new acronym in three months and we all pretend it is completely different. Tech likes doing that.
But the idea is useful. If AI becomes part of how people search, compare, and decide, companies need to think about the information AI can actually work with. Not just marketing copy, but the actual data behind products, services, categories, attributes, documents, and systems.
For a long time, storytelling helped companies win attention. I still think that matters. But AI needs more than a good sentence. It needs structure, context, consistency, and enough detail to compare, classify, explain, and recommend.
Product data makes this very concrete
Product data is a good place to see this problem, because it looks simple until you actually work with it. On paper, a product has a title, description, image, price, category, specifications, supplier details, maybe packaging data, maybe compliance fields. Fine.
Then the actual data arrives. One supplier sends Excel. Another sends CSV. Another has an API. Another sends PDFs. Images arrive separately. Units are inconsistent. Attribute names do not match. Categories are different. Required fields are missing. Someone still has to map, clean, enrich, validate, and publish everything.
This was already difficult when the goal was to get products into a PIM, ERP, webshop, or marketplace. Now add AI-driven search and recommendations on top of it. Suddenly, messy product data is not just an internal workflow problem. It becomes part of whether a product can be understood, compared, and recommended. Or ignored.
A good product can still disappear
Imagine someone asks an AI assistant: “What is the best jacket for cycling to work in rainy weather, under €150, available in size M, and made with recycled material?” That does not feel futuristic anymore. People already ask tools questions like this.
For AI to answer properly, it needs structured information: category, use case, price, availability, size, material, waterproof rating, sustainability information, and probably a few more attributes. If that information is missing, inconsistent, hidden in a PDF, written differently in every supplier file, or buried somewhere in free text, the product may not be understood well enough.
The product itself might be great. The data is not. That gap matters more when AI becomes part of the buying journey.
Bad data used to be easier to hide
Messy product data used to stay mostly inside the company. A supplier file breaks, someone fixes it. A marketplace export fails, someone checks it. A required field is missing, someone asks for it. A spreadsheet becomes “temporary” for three years.
The cost is real, but often hidden. It shows up as delays, rework, manual checks, failed imports, and a few people who know how everything actually works. Always dangerous, especially when they go on holiday.
But AI changes the pressure. If AI systems influence how products are discovered, compared, recommended, or excluded, unclear data becomes a visibility problem too. It is no longer only about internal efficiency. It is about whether your information is usable where decisions are being made.
Today you may be late for ChatGPT. Tomorrow you may be late for everything built on top of AI. That sounds dramatic, but markets often look stable until they do not.
Nokia also used to be a successful phone company.
The next advantage is understandable data
I do not think the future belongs only to the companies with the biggest AI budgets. I think it will also belong to companies whose data is easiest to understand, trust, and use.
Understandable data means clear attributes, consistent categories, validated fields, and workflows that can handle messy input before it spreads downstream. It means data that works for people, systems, and AI.
Yesterday, SEO helped companies get found. Tomorrow, structured data will help companies get chosen. And companies that wait too long may not disappear overnight. They may just slowly stop showing up in the places where decisions are being made and eventually become insignificant.
The author is Thorin, Founder, CEO and CTO of ProductFlight. View his LinkedIn profile.
This blog post is a contribution by ProductFlight. ProductFlight helps organizations with AI-driven product data transformation for modern retail. For more information, visit www.productflight.io or stop by the ProductFlight booth at Data Expo.