Ontology Process

Purpose

This design document explains the Opsfolio ontology process, why OWL is used, how the ontology is structured, how it is used in MCP-based AI systems, and how it helps define operational truth while reducing operational truth drift.

The short version is simple: the ontology gives Opsfolio a shared semantic model for evidence, assets, events, workflows, policies, incidents, metrics, and organizations. That shared model keeps humans, software, and AI aligned on what things mean and how they relate.

Why We Use an Ontology

Opsfolio operates across multiple operational domains: resource surveillance, security assurance, infrastructure monitoring, data governance, and organizational responsibility. Each domain produces data, but raw data alone does not create shared meaning.

The ontology provides that meaning by defining:

• the canonical concepts in the system
• the allowed relationships between those concepts
• the core constraints that prevent ambiguous interpretation
• the evidence model that ties observations to decisions

This turns operational data into a usable, reviewable truth model.

Why We Use OWL

We use OWL because it gives us a formal, machine-readable way to model operational semantics.

OWL is useful for Opsfolio because it supports:

• Classes for domain concepts such as Evidence, Resource, SecurityIncident, and DataWorkflow
• Subclass hierarchies so specialized concepts inherit shared meaning
• Object properties for relationships such as derivation, ownership, monitoring, and policy linkage
• Datatype properties for timestamps, values, and units
• Inverse properties for bidirectional reasoning
• Disjointness and restrictions for detecting model errors and preserving consistency
• Annotation vocabulary for human-readable definitions, examples, and scope notes

OWL is not being used here as an academic exercise. It is being used as the operational schema for meaning. The goal is not to model every possible implementation detail, but to define the stable semantics that should survive across databases, services, interfaces, and AI layers.

Ontology Design Principles

• Evidence First: Opsfolio is modeled as an evidence-centered system. Evidence is the unifying concept across domains.
• Shared Core, Specialized Domains: The model starts with a small abstract core and then specializes into operational domains.
• Operational Meaning Over Storage Layout: The ontology is designed around business meaning first, while still preserving links to implementation concepts.
• Traceability and Provenance: The model favors relationships that answer what was observed, where it came from, and who owns it.
• AI Grounding and Governance: The ontology is intended to be a durable semantic contract for AI assistants and MCP servers.

---

Ontology Process

The ontology process should be treated as a design and governance workflow rather than a one-time modeling task.

---

Current Ontology Processing Pipeline

The current implementation flow moves ontology content from source files into AI-accessible operational interfaces.

---

Ontology Structure

The current ontology is organized in a layered way.

Core abstract concepts

• Evidence: the central concept for verifiable operational truth
• Asset: something of organizational value under management
• Event: something that happens in time
• ProcessingEvent: an event that creates or processes evidence
• Observation: a time-specific measurement or reading

Domain areas

• Resource surveillance: Models digital resources, collections, and surveillance sources.
• Security assurance: Models endpoints, compliance policies, frameworks, and incidents.
• Infrastructure monitoring: Models network assets, infrastructure metrics, and thresholds.
• Data governance: Models workflows, data quality issues, and transformations.
• Organizational structure: Models organizations, people, roles, and monitoring behavior.

Relationship layer

The ontology connects these domains through relationships such as derivation and lineage, session tracking, policy linkage, and asset ownership. This layer turns a list of terms into an operational knowledge graph.

---

How the Ontology Is Used in MCPs

In MCP-based systems, the ontology acts as the semantic contract between tool outputs and AI reasoning.

1. MCP tools expose operational facts: MCP servers provide access to operational systems (evidence stores, policy engines, etc.).
2. The ontology provides the semantic frame: The ontology defines what those facts mean (e.g., distinguishing a Resource from ComplianceEvidence).
3. The AI uses ontology terms instead of ad hoc labels: Anchoring reasoning to the ontology improves consistency across prompts and sessions.
4. Cross-tool joining becomes reliable: The shared model lets AI connect data from multiple different MCP servers.
5. AI explanations become traceable: The AI can explain conclusions by pointing to specific ontology-backed evidence and provenance.

---

Ontology and Operational Truth

Operational truth requires a clear definition, evidence, provenance, time context, and relationships. The ontology helps define this by making those parts explicit and consistent.

Reducing Operational Truth Drift

The ontology reduces drift by providing:

• Canonical vocabulary: One shared language for core concepts.
• Typed relationships: Defining which connections are logically valid.
• Evidence grounding: Keeping conclusions tied to hard evidence.
• Provenance and lineage: Preserving how evidence was acquired.
• Cross-domain consistency: Connecting security, governance, and monitoring.

---

Guidance for AI and MCP Integration

To get the most value from the ontology, follow these rules:

• Return typed entities that map cleanly to ontology concepts.
• Include identifiers, provenance, and timestamps in tool outputs.
• Preserve links between evidence and the claims they support.
• Use ontology labels and definitions in prompts and explanations.
• Prefer canonical ontology terms over local database jargon.
• Version ontology changes and communicate impact to AI workflows.

---

Recommended Docs Site Structure

If this content is split into a docs.opsfolio.com design section:

1. Ontology Overview
2. Ontology Process and Governance
3. OWL Modeling Conventions
4. Ontology in MCP and AI Systems
5. Operational Truth and Truth Drift

Summary

The Opsfolio ontology is the semantic backbone for evidence-centered operations. Together with OWL and MCP, it creates a stable definition of operational truth and reduces the drift that appears when humans and AI operate from inconsistent concepts.