Ontology is having its moment

Haha I saw that coming! Obviously LLMs (and multi-modal world models to follow) need grounding. Ontologies, knowledge graphs, formal logic, these kinds of things can provide that.

The fads of AI are pretty hilarious. ML is so not cool. ML IS AMAZING! Ontologies are the Way! Ontologies are so passe. ONTOLOGIES ARE THE FUTURE.

Another thing I think will become obvious is that these generative models are intuition engines, not workflow engines. People are so busy trying to stuff all functionality of everything ever made into an LLM box that they are neglecting low hanging fruit in grounded execution pathways using workflows… that’s a fad that will give way. To be sure, a rigid/static workflow doesn’t have the flexibility to allow an agent to dynamically solve complex problems. But a hybrid solution with adaptive, machine-editable workflows allows agents to go much further by planning ahead and then reliably executing those plans over a long horizon.

Tony Seale said:

Ontology is having its moment. There was a time when we called it the “O word.” Nothing killed a conversation with business - or IT - faster than mentioning ontology. The rule was simple: deliver value, keep the ontology part quiet.

But things have changed.

Ontology is shaping up to be the buzzword of 2026. A big part of that is Palantir’s extraordinary rise - their entire Foundry platform is built on ontological modeling. Microsoft is now moving ontology into Fabric. The race is on.

Why now?

Because structured ontologies offer something generative AI desperately needs: grounding.

LLMs are creative but lack logical constraints. Ontologies provide the formal structure to anchor meaning and harmonize wildly different data sources into a coherent semantic layer. When you need precision over plausibility, ontologies are one of the most reliable ways to keep AI from hallucinating.

But what actually is an ontology?

The idea traces back 2,000 years to Aristotle, who tried to formalize how we perceive reality. Ontology today is that same craft: defining the concepts and relationships that shape our world in a logically rigorous way.

Fast forward to the early web. Tim Berners-Lee envisioned the Semantic Web - an internet where data, not just documents, was linked. Give every ontological concept a unique identity, he argued, and meaning becomes a first-class citizen of the web.

Google later operationalized part of this vision with Schema.org, creating a shared vocabulary to help search engines understand the public internet.

But the current frontier isn’t the public web - it’s the enterprise.

Companies want AI agents that can reason about their internal data - their customers, assets, processes. They need semantic clarity applied to their own messy, private reality. Disambiguating meaning isn’t just useful anymore. It’s essential.

Here’s the problem: you can’t outsource your ontology. Enterprises need their own internal version of schema.org- fully owned, built on open standards. Anything less means IP leakage and vendor lock-in. In the age of AI, your ontology is a key part of your competitive moat.

So, ontology is back. And this time it’s not theory - it’s critical infrastructure. ⭕ Your Schema.org: https://lnkd.in/eumPB3Hj ⭕ Your Ontology Your IP: https://lnkd.in/ersgR-Df