AI systems need more than access to data — they need explicit definitions of business concepts, relationships, and metrics. How do we design a semantic layer that supports both analytical workloads and AI agents without creating competing interpretations of the same business? This session explores that architecture through Microsoft Fabric IQ, examining how Power BI semantic models and ontologies complement each other. We will discuss the distinction between analytical models and business ontologies, how to map entities and relationships to underlying data, and how shared semantics can ground natural-language queries and agent interactions. The focus is on practical design decisions: reusing existing semantic models, defining entity boundaries and relationships, maintaining consistent business terminology, and understanding where semantic context helps—and where it cannot replace data quality, access controls, or validation. Attendees will leave with a framework for evolving a reporting-oriented semantic layer into a foundation for AI-enabled applications.