--- title: "Data foundations: the step that comes before AI" description: "\"We want AI on our technical documentation. Where is your documentation? In binders.\" The data foundation comes before AI, and that is good news." url: "https://gilabs.fr/en/blog/socle-data-avant-ia/" lang: "en" author: "Yann Gilliot" published: "2026-07-10" tags: ["Data maturity", "AI Transformation", "SMEs", "Diagnostic"] --- # Your problem is not (yet) an AI problem By [Yann Gilliot](https://gilabs.fr/en/about/) ยท published on 2026-07-10 "We want AI on our technical documentation. Where is your documentation? In binders." The data foundation comes before AI, and that is good news. ## The short version - AI plugs into a foundation: digitised documents, structured data, a source of truth. Without that foundation there is no useful AI. - The honest diagnosis is often that the problem is not yet an AI problem, and that is good news: the foundation is simpler, cheaper and faster to build. - A test you can run yourself: if an AI had to answer your business questions tomorrow, where would it go looking? > "We want to put AI on our technical documentation. > Where is your documentation? > In binders." This exchange, from a recent conversation with an industrial company, is not an isolated case. It is actually a situation I run into regularly with clients. ## The real diagnosis is not the one you'd expect Many business leaders tell me the same thing: "we are behind on AI." Then we look together at what is actually happening on the ground: the technical documentation exists on paper, data is scattered across disconnected spreadsheets, and the version that counts as authoritative lives, in reality, inside one person's head. The real diagnosis, in this kind of situation, is not "you are behind on AI." It is: your problem is not yet an AI problem. ## AI does not plug into paper An AI system, however capable, plugs into a foundation: digitized documents, structured data, a source of truth that is unambiguous. Without that foundation, there is no viable AI. Or worse: there is an AI anyway, in the form of a chatbot bolted onto the existing disorder, answering confidently based on wrong or incomplete data. [An AI that is confident and completely wrong](https://gilabs.fr/en/blog/agent-n8n-prospection/) is something I ran into on my own agent. ## Good news, not bad news The good news in this diagnosis is that building this foundation is a noticeably simpler, cheaper and faster project than an AI project. And it pays off from day one, independent of any AI: finding a document in seconds, knowing with certainty which version of a plan is authoritative, no longer re-keying information that was already entered elsewhere. ## Why data maturity weighs as much as process pain In [a 30-day AI diagnostic](https://gilabs.fr/en/blog/ce-que-le-diagnostic-demande-aux-equipes/), data maturity weighs as much as the pain of the process itself. And the honest recommendation is sometimes: no AI just yet. The foundation first. AI will plug into it afterward, and only then will it hold up over time. ## A simple test you can run yourself One question is enough to place yourself: if an AI had to answer your business questions tomorrow, where would it go looking for answers? If the honest answer is "in a binder" or "you'd have to ask so-and-so," you have just identified your step one, before AI even enters the conversation. The [30-day AI diagnostic](https://gilabs.fr/en/services/ai-diagnostic/) says so too, when that is the case. Topics: Data maturity, AI Transformation, SMEs, Diagnostic ## Take it further Is your data ready, or still sitting in binders? Before you spend a euro on AI, you know what it will return. In 30 days: your processes mapped, the real end-to-end delay of each one, a return in hard euros per priority case and a payback measured in months. My method has three steps, map it, prioritise it, then automate it, in that order: automating a cluttered process does not speed it up, it multiplies it. That is why part of the gain I quantify for you cannot be bought, least of all from me. Quantified by the person who will build it. [30-day AI diagnostic](https://gilabs.fr/en/services/ai-diagnostic/) **Book a 30-minute call**: [Calendly](https://calendly.com/yann-gilabs/30min)