One meaning for every record
Before an organisation can trust its AI, it has to agree on what its own data means. The data dictionary is the least glamorous and most important AI investment.
Read articleThe word sounds academic. The idea is practical: describe your world once, precisely, so that every system and every model understands it the same way.

Ontology is one of those words that makes executives reach for their phones. It shouldn’t. Stripped of jargon, an ontology is simply a shared, precise description of the things an organisation cares about and how they relate to each other.
Databases store rows and columns. Decision-makers think in terms of real things: a vehicle, a depot, a route, a driver, a delivery. They also think in relationships: this driver is assigned to this vehicle, which serves these depots, which supply these customers. An ontology captures those things and relationships explicitly, so that software can reason about them the way people do.
A dictionary tells you what a word means. An ontology tells you how everything fits together.
The best ontologies are built with the people who run operations. They know which distinctions matter and which don’t. Engineers then make that knowledge machine-readable. When done well, the ontology outlives any individual application, and it becomes one of the most valuable assets an institution owns.
Choose one domain where decisions are frequent and costly. Name the things, the relationships and the rules. Connect them to the data you already have. You will be surprised how quickly the conversation shifts from “where is the data?” to “what should we do about it?”
Want to explore how this applies to your organisation? We share our approach with qualified organisations in a private briefing.

Before an organisation can trust its AI, it has to agree on what its own data means. The data dictionary is the least glamorous and most important AI investment.
Read article
Twenty months of writing about data, AI and sovereignty point to one idea: an intelligence layer that turns what an enterprise already owns into foresight, under its own control.
Read article
AI adds new things to protect: training data, models, prompts and, increasingly, the actions systems can take. Zero-trust principles apply to all of them.
Read article