The Ontology & Type Definitions capability governs how knowledge is structured and validated within the Fuse Knowledge Graph using standardized node and link types.
An ontology defines the types of knowledge entities (e.g., documents, people, policies) and the relationships between them (e.g., authored_by, references, owned_by).
It ensures consistency, interoperability, and explainability across knowledge sources, extraction pipelines, AI prompts, and workflow logic.
Component | Description |
---|---|
NodeType |
Defines a category of knowledge (e.g., Document , Person , Concept ) |
LinkType |
Defines valid relationships between nodes (e.g., references , belongs_to ) |
SchemaSet |
A collection of NodeTypes and LinkTypes for a given domain or scope |
Ontologies may be:
Scopes allow reuse while preserving contextual relevance.
Capability | Description |
---|---|
Ensure graph entities carry known meaning and behavior | |
Add new types or link rules as domains evolve | |
Enforce allowable connections and detect conflicts | |
Map external data to Fuse types for consistency | |
Filter or report based on ontology-aligned structure |
PolicyDocument
node may only be linked to a Department
node via owned_by
Project
node may reference
another Project
or Initiative
, but not a File
Term
node can only appear in a defined_in
relationship with a GlossaryPage
These rules guide graph construction, extraction pipelines, and reasoning workflows.
Next: Prompt Context Builder — Learn how Fuse composes and injects relevant, access-controlled knowledge into AI prompts using embeddings, graph traversal, and scoped logic.