Enterprise Ontologies: Why They Matter Now
At UGROUND, we began researching Ontologies and Model Engineering back in 2008–2009, as a result of a project in the field of hospital process management. From that moment on, ontologies became the core design principle of our company, and we built an entire technological universe around them. For years, few people talked about this subject; ontological matters were generally confined to academic circles, with hardly any business application.
In recent years, however, there has been a huge shift: now everyone is writing about ontologies. Take a look around Medium or LinkedIn and you’ll find plenty of material being published. They’ve become fashionable! This has led us to look back a little and understand where all this is coming from.
What has happened for such abstract topics to become part of everyday discussion?
In my view, there are three clear culprits. The first is LLMs. We all know that LLMs are marvelous — except when they “go off the rails.” And that happens because, in large contexts, the ocean of words swallows up meaning, and everything loses sense. To give LLMs reliability, experts have discovered that ontologies work wonders. Instead of feeding LLMs mountains of text, we give them knowledge graphs, and everything starts working beautifully. So LLMs and ontologies have become the best of friends.
The other ally is digital twins. The idea of the digital twin was born in industrial applications, to describe things like parts, machines, or factories, and to enable simulation, prediction, and optimization. This concept is rapidly expanding into other domains such as digital twins of organizations (business management), of people (healthcare), and in defense or security (complex scenarios), among others.
But… what do ontologies have to do with this? A digital twin aims to faithfully represent a reality: a factory, a company, a person, or a crisis scenario — but the twin is not “programmed” for a specific use. The twin’s configuration information is descriptive, something along the lines of “you will operate in this context, which has these rules of the game, to achieve these results.” The very technique that allows us to model digital twins of any kind is ontologies — and there we have the second pillar of their current relevance.
And the third is, curiously enough, an American company that became very trendy last year: Palantir. This company has been creating top-tier technologies in military intelligence since 2001 (its core model). It is now transferring that know-how to enterprises in the form of digital twins. And its representation model is, precisely, ontologies.
In short, ontologies have become an imperative need in the new model of business management and technological innovation. That said, developing large-scale architectures based on ontologies is far from trivial, and there is still a long road ahead. In 2017 we obtained a US patent on ontology-based systems engineering, which is the core of our solution architecture and allows us to deliver very powerful results in new areas such as enterprise AI and next-generation human interfaces for very large business models.
Welcome to the new world!
Fdo.: Alfonso Diez
CEO