semantic-model-consumption

Call the ExecuteQuery MCP tool with the semantic model's artifactId and an array of 1-4 DAX queries, each containing a single EVALUATE statement. Find the artifactId first using the DiscoverArtifacts MCP tool with the model name.

Use DAX INFO.VIEW functions: INFO.VIEW.TABLES() for tables, INFO.VIEW.COLUMNS() for columns, INFO.VIEW.MEASURES() for measures, and INFO.VIEW.RELATIONSHIPS() for relationships. Narrow results with SELECTCOLUMNS and FILTER to reduce output volume.

INFO.VIEW.* functions are available to any user with read access and are preferred for first-pass metadata discovery. Many INFO.* functions expose deeper diagnostics and storage internals but may require elevated semantic model permissions.

The FabricIQ MCP server may not be registered, not loaded, or the client session has stale tool metadata. Verify the active MCP server and tool inventory and confirm ExecuteQuery is exposed before retrying.

No. This skill is strictly read-only, limited to metadata discovery and analytical DAX queries. For creating, deploying, or managing semantic model definitions, use the semantic-model-authoring skill instead.

Role members are assigned at the service level through Entra ID after deployment, not stored in the model definition, so DAX INFO functions cannot reliably surface them. Use the Fabric service role membership management documentation instead.