This paper presents the development and initial testing of an ontology-based AI agent built on the Value-Conflict Mapping (VCM) method—an approach for detecting and resolving contradictions in complex systems. VCM begins by analyzing the relationships between stakeholder requirements and system component properties, inverting property values to reveal trade-offs and linking them to alternative stakeholder needs. To support this process, we developed a set of ontologies covering stakeholder requirements, system requirements, and property contradictions, which serve as the semantic backbone for the AI agent. The agent processes system documentation, reasons over stakeholder–system correspondences and systematically identifies contradictions through ontology-grounded knowledge graphs. The paper outlines the agent’s modular architecture and illustrates its application in VCM analysis. Finally, it discusses the broader applicability of this approach, highlighting how ontology-based AI agents can extend TRIZ methodologies to other models and tools, thereby enhancing system design, contradiction detection, and innovation support.