Uploaded June 2026 | Updated November 2026, 3 days ago
A semantic model is a curated unit of Dataverse tables enriched with system-inferred signals and human-curated input, giving agents the shared understanding they need to reason over your data and power Dataverse agentic search. Each semantic model combines three building blocks.
Data scope and semantic indexing — Defines the set of Dataverse tables covered by the semantic model and enables semantic indexing over them, forming the foundation for all agent reasoning over a certain agent scenario:
* System-inferred signals — Automatically extracted from your existing Dataverse metadata, including table relationships, views, forms, description summaries, and sample data
* Manual input (Glossary) — Domain-specific terms and definitions added by makers to close gaps that system signals cannot infer, such as business-specific acronyms or organization-specific naming conventions
Together, these building blocks allow agents to understand not just what your data contains, but what it means — enabling more accurate, context-aware responses.
Learn more about Dataverse Intelligence Layer, read the latest blog: https://aka.ms/DataverseJun2026
Learn more about semantic models: https://aka.ms/DataverseSemanticModel
Speaker: Jason Huang, Senior Product Manager, Microsoft
A semantic model is a curated unit of Dataverse tables enriched with system-inferred signals and human-curated input, giving agents the shared understanding they need to reason over your data and power Dataverse agentic search. Each semantic model combines three building blocks.
Data scope and semantic indexing — Defines the set of Dataverse tables covered by the semantic model and enables semantic indexing over them, forming the foundation for all agent reasoning over a certain agent scenario:
* System-inferred signals — Automatically extracted from your existing Dataverse metadata, including table relationships, views, forms, description summaries, and sample data
* Manual input (Glossary) — Domain-specific terms and definitions added by makers to close gaps that system signals cannot infer, such as business-specific acronyms or organization-specific naming conventions
Together, these building blocks allow agents to understand not just what your data contains, but what it means — enabling more accurate, context-aware responses.
Learn more about Dataverse Intelligence Layer, read the latest blog: https://aka.ms/DataverseJun2026
Learn more about semantic models: https://aka.ms/DataverseSemanticModel
Speaker: Jason Huang, Senior Product Manager, Microsoft










