This paper proposes a conceptual ontology-driven governance layer for AIOps systems, addressing the Semantic Governance Gap — the structural absence of machine-processable semantic traceability in current AIOps architectures.
Building upon principles from ontology engineering, temporal-semantic modeling in knowledge graphs, and recent work on Knowledge Graph integration for log anomaly detection, the paper introduces a formal semantic framework that structures the lifecycle of operational decisions — from incident detection through recommendation and feedback.
The proposed model enables explainability by design, transforming AIOps from systems with post-hoc statistical interpretations into structurally transparent decision architectures, with structural alignment to EU AI Act requirements (Articles 11–17). Empirical validation is identified as future work.