Concept

Ontology

A structured model of concepts, relationships, constraints, and meaning within a domain.

SemanticsLevel 4

Position in the Trusted Data Framework

Ontology

Living Graph View

Ontology in the concept network

Drag nodes to adjust the view, double-click a node to expand its neighbourhood, and right-click a node to hide it while preserving the exploration context.

Loading concept graph

Relationship Matrix

Ontology as a network node

Connected concept
Relationship

Builds on real-world objects, events, places, and responsibilities before turning them into formal concepts.

Provides shared meaning that the semantic layer can operationalise across data products, APIs, dashboards, and AI applications.

Defines graph schema, relationship semantics, and constraints.

Gives AI systems governed domain meaning for retrieval, interpretation, and explanation.

Depends on persistent entities and identifiers so concepts can be attached to stable things rather than ambiguous records.

Constrains valid representations and manages the impact of semantic change.

Concept Understanding

How to recognise Ontology

Without Ontology
Duplicated meaningInconsistent APIsWeak semantic layerPoor AI reasoningFragile knowledge graph

Without ontology, teams may still store data, but they cannot reliably share meaning across systems, decisions, and AI applications.

Before OntologyBuildingFacilityAssetPropertyInfrastructure
After OntologyBuildingPhysical AssetFacilityOperational AssetAI-readable meaning

Ontology turns competing labels into explicit concepts and relationships that people, systems, and AI can interpret consistently.

Why it is not

Database Schema

A schema describes how data is stored. An ontology describes what the domain means.

ER Diagram

An ER diagram models database entities and relationships. An ontology models business concepts, semantics, and constraints across systems.

Data Dictionary

A data dictionary documents fields. An ontology connects concepts, relationships, rules, and meaning.

Traditional BIBusiness GlossaryMetadataOntologyKnowledge GraphWorld Model

Concept Relationships

How Ontology works with other concepts

Builds on real-world objects, events, places, and responsibilities before turning them into formal concepts.

Example

A field inspection maps a real bridge component; the ontology defines whether it is an Asset, Component, Location, or Observation.

Consulting question

Where does the real-world boundary of each concept begin and end?

Provides shared meaning that the semantic layer can operationalise across data products, APIs, dashboards, and AI applications.

Example

A shared Asset ontology becomes consistent asset definitions for reporting, semantic APIs, and AI copilots.

Consulting question

Which domain definitions must remain consistent across systems and teams?

Defines graph schema, relationship semantics, and constraints.

Example

Asset, Location, Event, Document, and Risk classes become graph nodes and typed relationships.

Consulting question

Which relationships need to be explicit before the organisation can reason across data?

SupportsAI Reasoning

Gives AI systems governed domain meaning for retrieval, interpretation, and explanation.

Example

A RAG assistant can distinguish an asset owner, maintainer, operator, and regulator instead of treating them as generic organisations.

Consulting question

What concepts must AI understand correctly before recommendations can be trusted?

Depends onIdentity

Depends on persistent entities and identifiers so concepts can be attached to stable things rather than ambiguous records.

Example

A building, asset, facility, and operational system can be distinguished only when identity rules are clear.

Consulting question

Which entities need persistent identity before semantic modelling can be trusted?

Constrains valid representations and manages the impact of semantic change.

Example

Changing the definition of Critical Asset must trigger review across dashboards, risk models, and knowledge graph mappings.

Consulting question

Who can approve semantic change, and how is the impact assessed?

Definition

An ontology defines the core concepts in a domain and the relationships between them. It gives systems and people a shared model of meaning.

Why It Matters

AI systems need more than data fields. They need context, constraints, and relationships. Ontologies help organisations make data interoperable, explainable, and reusable across systems.

Role in the Trusted Data Framework

Ontology sits mainly in the Semantics layer. It connects information to knowledge by making domain meaning explicit and governable.

Practical Examples

  • Defining common asset, location, event, and organisation concepts for a digital twin.
  • Mapping inconsistent source-system terms to a shared business vocabulary.
  • Supporting ontology-driven data integration for analytics and AI workflows.

Case Evidence

GBlocks' ontology (semantic/ontology.yaml) is deliberately lightweight -- 19 concepts (grown from an initial 15) and a small closed vocabulary of relationship types (extended from 10 to 12 when two new real relationship shapes appeared) -- but it encodes three things worth noticing. First, confidence is part of the ontology itself: the Transaction refers_to Property relationship states outright that only 7% resolve to an exact matched identity, the rest fall back to a coarser relationship, so a consumer knows how much to trust it before using it. Second, governance facts live inside the model, not in separate documentation -- the PhysicalFeature concept's own definition notes that two of its three source tables are billed Premium products, discovered only after the fact via a usage check, not assumed at design time. Third, ontology discipline caught a duplicate before it ever entered the model: while adding four new sources, a fifth candidate dataset was recognised, via schema inspection, as a republication of a source already loaded under a different publisher label -- and was never ingested. Recognising when two differently-labelled sources describe the same real-world concept is as much ontology work as naming the concepts in the first place.

Consulting Questions

  • Which business concepts must be shared across teams and systems?
  • Where do source systems use different names for the same thing?
  • Which relationships are essential for decision-making?
  • Who owns and governs semantic change?
  • How will the ontology be validated against real data?