Docs · Semantic governance
Semantic governance administration
Learn how SentinelAI uses Operations-hosted taxonomy, ontology, and graph workflows to connect use cases, models, datasets, controls, and governance documents.
Overview
Semantic governance gives operators a shared administration layer for controlled taxonomy terms, ontology relationships, and graph-backed exploration. It is designed to make cross-object governance clearer and more explainable without presenting SentinelAI as a separate graph-database product.
This page is part of the public SentinelAI documentation layer. It is meant to accelerate orientation and evaluation while staying aligned with the product’s governance-focused positioning and messaging guardrails.
Taxonomy administration
Operators can use the Operations-hosted Semantic Admin workspace to create taxonomies, maintain hierarchical terms, define aliases, and control activation state so governance classifications remain reusable across product workflows.
- Create taxonomies with durable keys, display names, owner context, and version-aware change discipline.
- Add and edit terms with parent-child hierarchy, aliases, descriptions, and active or inactive status instead of relying on unmanaged free text.
- Use shared taxonomies for lifecycle states, evidence categories, jurisdictions, sensitivity labels, and other governance classifications that need to stay consistent across teams.
Ontology and relationship administration
Entity types and relationship types help teams define what can connect to what, which makes governance relationships easier to manage over time.
- Review and manage ontology entity types for governed objects such as models, use cases, datasets, controls, documents, and vendors.
- Configure relationship types with source and target rules, reverse labels, directionality, multiplicity, and metadata hints so the graph stays explainable.
- Treat semantic keys as stable identifiers and use names, descriptions, and metadata for change control instead of renaming live relationship codes casually.
Graph console and impact exploration
The standalone graph console supports neighborhood review, impact summaries, framework mapping, propagation views, and saved-query entry points from one operational surface.
- Start from a known object UUID or use a saved semantic query template to jump into a connected governance investigation quickly.
- Inspect incoming and outgoing relationships, downstream impact, framework mapping, and propagation summaries without opening records one by one.
- Use graph-backed exploration when reviewers need to explain why a model, dataset, use case, control, or document is connected to a broader governance decision.