# Data Is Not the New Oil. It Is the Foundation > Artificial intelligence is only as good as the information it receives. Why data isn't a byproduct but the foundation of every successful AI implementation. *Published: 2026-04-05* *Source: https://vellmerk.ai/en/blog/daten-als-fundament* For years we've been hearing the same phrase: "Data is the new oil." A nice metaphor, but it's misleading. Oil is a raw material that you extract, refine, and burn. Data is something different. Data is the foundation on which everything stands. Think of artificial intelligence as a house. Everyone talks about the architecture, the facade, the smart features. But nobody asks: How solid is the foundation? Across multiple client projects, Vellmerk.ai sees the same pattern again and again: most AI projects fail not because of the technology, but because of the information base. ## The Information Foundation What exactly do we mean by foundation? It is everything that an AI has at its disposal to solve a task. In practice, we encounter various terms for this: **Prompt Engineering**: the art of asking an AI the right question and providing the right context. Sounds trivial, but it isn't. The difference between a mediocre and an excellent AI response almost always lies in the prompt. **Context Engineering**: the next step: not just optimizing the question, but systematically shaping the entire information context that the AI receives. Which documents, which rules, which examples are included? **Data Management**, when we take yet another step back: How is the company's data structured, maintained, and accessible? Here we're talking about data quality, data architecture, and governance: the basis on which everything else is built. All these terms fundamentally describe the same thing: different levels of the same information foundation. And all of them need to be right for AI to work reliably. ## The Higher the House, the Deeper the Foundation A simple chatbot that answers FAQs needs a manageable foundation: a clean knowledge base, clear wording, done. But an AI agent that independently manages business processes, makes decisions, and interacts with external systems? That needs a massive foundation. And here lies the problem: Many companies want the penthouse but have the foundation for a garden shed. The consequences are well known, under various names: **Hallucinations**: the AI invents information because it lacks the right data. **Errors in automated processes**: the workflow runs, but the results are wrong because the input data is incomplete or outdated. **Inconsistent answers**: the same question delivers different results because the context is not reproducible. These are usually not technology problems. They are data problems. ## The Uncomfortable Mirror Now let's be honest: When we talk about "artificial intelligence," it sounds like an independently thinking entity. But it isn't. What we call AI is, at its core, highly complex stochastic data processing. No consciousness, no intuition, mathematics based on probabilities. And that means: When the result isn't right, in most cases it's not the computation that's wrong, it's what we fed into it. > "When AI makes mistakes, we need to be honest: Most of the time it's not the AI, it's us. The data we gave it. Or didn't give it.", Thorsten Vellmerk The old principle of "garbage in, garbage out" is as relevant as ever. The only difference is that the consequences are bigger today, because AI systems are being deployed in increasingly critical processes. ## From Foundation to Flying Use Case The good news: The foundation can be built. In its consulting, Vellmerk.ai therefore doesn't start with the question "Which AI tool should we buy?" but with: "What does your data foundation look like? And what does your use case really need?" Sometimes it's enough to properly structure existing documents. Sometimes a data architecture overhaul is needed. And sometimes it turns out that the desired use case doesn't fail because of the AI, but because of missing processes upstream. Figuring out exactly that, which foundation you need so your specific use case can fly, that's what Vellmerk.ai does every day. > "Data is not the new oil. Oil gets burned. Data is the foundation, and on a good foundation, you can build as high as you want.", Thorsten Vellmerk ## Conclusion Before you think about AI tools, agents, or automation: Look at your foundation. Is your data clean, accessible, and complete? Is the context your AI receives really what it needs? Is the basis on which everything is built solid? If you're not sure, that's exactly what we're here for. **Get in touch**, and Vellmerk.ai helps you build the right foundation for your AI initiatives. ## About Vellmerk.ai Vellmerk.ai is an AI consultancy (Danish ApS) founded by Thorsten Vellmerk. Drawing on 20+ years of process and IT experience and several years of hands-on AI consulting, Vellmerk.ai helps SMEs and public administration adopt AI in a practical, sovereign way, from strategy to local, on-premise-ready implementation. Proven across multiple client projects. [Book an initial consultation](/en/contact). ## Frequently asked questions ### Why is data more important than the AI model? Because even the best model is only as good as the data it works on. Models today are largely interchangeable building blocks; the real value sits in your data and its quality. Poor, incomplete or unstructured data reliably leads to poor results, no matter how powerful the model. ### What does data quality mean for AI in concrete terms? Data quality means the data is available, accessible, consistent, current and sufficiently structured for the intended use case. It also means it is clear what the data means and where it comes from. For AI it is not the largest volume of data that counts, but the data basis that fits the task and is reliable. ### Does all data have to be perfect before AI can start? No, and waiting for that would be a mistake. You do not need a perfect, company-wide data basis, but a good enough data foundation for the specific first use case. Data work and AI adoption run best in parallel: you start focused, learn, and keep improving the data basis where it pays off. ### What is the first step towards an AI-ready data basis? Getting an honest overview: what data exists, where it sits, in what quality, and who may use it. Only from this inventory does it become clear which use case will realistically work first. Often the biggest lever is not a new tool, but order and access to data you already have.