*This article is a restructured version of Hydraft®’s consulting assets.
Recently, in the AI field, Ontology AI has been attracting more attention due to the limitations of RAG (Retrieval-Augmented Generation). Since most of my clients are actively utilizing RAG at their points of business interaction, I have prepared this brief overview to understand the market and technological changes in advance.
Al though the domestic market has not yet reached a level demanding Ontology AI, companies that aim to lead the market need to proactively understand the concept of Ontology, prepare a structurally sound data infrastructure, and plan long-term roadmaps.
RAG, which has been in the spotlight in recent years, generates answers by simply retrieving documents. As such, it has limitations when document quality is low, data is insufficient, or structures are incomplete, making it difficult to provide accurate responses.
Furthermore, RAG references each document fragment independently, which means it cannot fully understand semantic relationships or structural contexts between different pieces of information. This results in limitations for complex logical reasoning and generating consistent answers.
Unlike RAG, Ontology AI does not stop at document search or pattern recognition. Instead, it understands and connects semantic relationships and structural contexts between data. By structuring relationships between data in the form of a knowledge graph, it can create a logical and hierarchical knowledge system.
Ontology AI enables complex conceptual reasoning, and even with incomplete information, it can interpret context and derive meaningful conclusions. This makes Ontology AI increasingly important in fields such as education and industry, where results based on complex data are required.
To understand Ontology AI more deeply, it is first necessary to briefly examine two core concepts of an ontological system: entities (beings) and relationships. The structure of entities and relationships is by no means a new concept. It has been studied for centuries, and the first systematic discipline to investigate and establish these ideas was Aristotle’s “First Philosophy,” or Metaphysics. Metaphysics is also the theoretical framework I frequently refer to in my professional work when explaining the concept of brand substance.
As a central branch of metaphysics, Ontology is the discipline that directly investigates Being itself—that is, what it means for something to exist. To explain the relationship between these disciplines in greater detail, metaphysics can be understood as the broader field concerned with the most fundamental principles and structures of reality, including Being itself. Within this broader framework, ontology examines how individual entities (beings) exist and what kinds of relationships they have with one another.
Substance Theory (substance, ousia), one of the key theories within ontology, is particularly relevant to brand theory. It can be understood as the study of the fundamental principle that allows an entity to remain what it is despite undergoing change.
Accordingly, the philosophical hierarchy between metaphysics and ontology introduced above can be summarized in the following order:
From this perspective, the Ontology discussed in the IT field can be understood as a structural implementation of ontological thinking from metaphysics at the data infrastructure layer.
In other words, it defines what exists (Being) within a given data environment and systematically structures the relationships among those entities (beings).
Ontological thinking itself predates Aristotle. However, what we recognize today as a systematic philosophical discipline of Ontology is generally understood to have developed from the tradition of inquiry established through Aristotle’s Metaphysics. It is important to note, however, that Aristotle did not distinguish Ontology and Substance Theory as separate disciplines in the way we do today. These distinctions emerged much later, particularly through philosophers from the seventeenth century onward.
For reference, substance (ousia) can be understood in Aristotelian terms as the unity of matter and form.
Traditional data systems often struggle to understand context across a sequence such as “customer → order → product.” Instead, they primarily process surface-level text or attribute patterns associated with each individual stage.
In other words, it is difficult for such systems to understand the semantic connections and relationships between customer data and order data, for example.
An Ontology system, by contrast, defines the meaning of data at the level of entities and relationships, and connects them through a structured semantic framework.
It is, in a sense, like having a map of relationships within the data.
Because these relationships and meanings are explicitly structured, an AI system built on an Ontology does more than simply retrieve data and return results. It can use relationships and context to perform domain-specific, coherent reasoning and derive new meaning from the underlying structure.
To better understand the characteristics of Ontology, consider a simple example based on a Hospital Management System (HMS).
Within an Ontology framework, a patient's statement such as “I have a headache” is not treated merely as a piece of text. It can instead be represented as a structured semantic relationship:
has symptom → headache
Based on these logical rules connecting entities and relationships, AI can perform semantic inference — for example, reasoning about what treatment a patient receives, which physician is involved, and what medication may be prescribed.
Structurally defining and connecting entities and relationships within an Ontology system is essentially the process of projecting the structure of the real world into the data layer.
This approach is also closely connected to constructivism and to the use-case-oriented system design of MSA (Microservice Architecture) discussed in the previous article.
As mentioned earlier, Ontology-based AI is attracting attention as a potential next-generation approach beyond RAG, which often remains focused on retrieving fragmented pieces of information.
If RAG primarily answers “What are you asking?”, Ontology seeks to explore “Why are you asking it?”
In other words, an Ontology-based approach goes beyond simply providing information. It seeks to understand the underlying structure of knowledge and derive meaningful conclusions from it.
If what we ultimately seek from AI is not merely search and summarization, but rather contextual understanding and the generation of new meaning, then an Ontology built around the structure of entities and relationships may inevitably become an increasingly important foundation beyond conventional RAG.
There is currently one company that has developed one of the world's most sophisticated implementations of an Ontology system: Palantir, led by Peter Thiel and Alex Karp.
Palantir applies the ontological way of thinking rooted in Aristotle's Metaphysics directly to the business domain, and calls its approach the Palantir Ontology.
Rather than simply enabling data retrieval, the system is designed to understand an organization's knowledge structure, integrate it with operational processes, and use that structure to simulate and extend real-world operations. In this sense, Palantir aims to build an AI-enabled end-to-end digital twin of an organization.
By integrating an organization's data and processes through a common semantic structure, Palantir's approach enables forms of reasoning and simulation that are difficult to achieve through conventional RAG alone.
In a future article, I will take a closer look at the story behind Palantir, why it has attracted significant attention in the investment market, and what makes its business model distinctive.
Thank you.
© 2025 Hydraft®
All rights to this content are reserved by Hydraft® and are protected under copyright law. Unauthorized reproduction and redistribution are strictly prohibited.
