Created on 2025-05-25 17:15
Published on 2025-05-25 17:25
For decades, the dream of business leaders and technologists alike has remained remarkably consistent: describe a business model clearly and orchestrate its execution seamlessly—with the help of software. From spreadsheets and databases to APIs and SaaS, and now through GenAI and intelligent agents, we've been chasing that dream.
But to truly realize it, we must confront two words many in tech have long sidestepped: Philosophy and Ontology.
Because in the end, philosophy will eat AI—not by competing with it, but by abstracting, absorbing, and elevating it.
Every technological leap—from the personal computing era to the Web, to today’s GenAI—has aimed to simplify and automate business logic. But most solutions fall short. They leave enterprises entangled in fragmented systems, opaque decision-making, and automation that only loosely mirrors the real-world business it’s meant to serve.
Worse still, the result is often not transformation, but accumulation—yet another layer of compounding technical debt, piled on top of yesterday’s quick fixes and tomorrow’s integration headaches.
The reason? We’ve lacked a shared, machine-readable way to express the meaning behind what we do.
Just as the Web abstracted away DNS by allowing us to focus on what we’re looking for (documents, services) rather than where they reside (machine network addresses), ontologies will do the same for GenAI.
Ontologies provide the semantic scaffolding—the structure of meaning—that allows software to understand rather than just respond. They define the nature of relationships between concepts in a way that both machines and humans can interpret. When combined with GenAI, they unlock a level of explainable, reusable, and orchestrated automation that reflects the actual logic of the business.
By integrating logic and relationship type semantics through what's expressed in an ontology, we eliminate the barriers between data, decision-making, and execution. This allows leaders to:
Model enterprise intent as software
Compose agents that reflect actual business roles and knowledge
Orchestrate workflows based on meaning, not syntax
In short, we move from coding instructions to expressing intent.
Today, this vision is becoming reality through the emergence of protocols like the Model Context Protocol (MCP). MCP enables the creation of modular, composable agents—each linked to the tools, data, and logic that define specific business functions and capabilities.
Working in tandem, the Agent-to-Agent (A2A) protocol supports dynamic collaboration between these agents—empowering them to coordinate, reason, and adapt in real time. The outcome is no longer static (an draconian) software, but living workflows that faithfully model the intentions and operations of the business itself.
These aren’t just smarter systems—they are systems that think like the business thinks. In doing so, they elevate the enterprise architecture stack: above systems-of-record and systems-of-engagement, and even above systems-of-intelligence, we now introduce systems-of-thought—a new layer where meaning, logic, and intent drive automation.
What we're building is not just automation, but philosophical software: systems that encode how we think, reason, and act as enterprises. It’s no longer just AI that augments our work—it’s philosophy made executable through AI.
And that’s why, in the long arc of technology, philosophy will eat AI.
Because to truly transform enterprise software, we must move beyond tools that mimic intelligence—and start building systems that understand and reflect it through a symbiotic relationship between human supervisors and the agents they oversee.
Want to go deeper? Philosophy Eats AI: What Leaders Should Know