If you want a Power BI semantic model to be reusable by multiple reports without an enterprise data warehouse, which component should you create?

• A. Linked table
• B. Shared dataset
• C. Staging dataflow
• D. Transformation dataflow

Answer: (B)

Centralizing the semantic model as a shared dataset in the Power BI service lets multiple reports connect to the same data model. This means the definitions of tables, relationships, measures, and security rules live once and are reused everywhere, ensuring consistency and reducing duplication. You publish the dataset to a workspace and then reports can connect to it (either live or by import) and refresh together, without needing an enterprise data warehouse. The other options don’t provide this single, reusable semantic layer across several reports: a linked table stays within an individual report’s model, and staging or transformation dataflows are ETL steps that prepare data but don’t by themselves serve as the centralized model that multiple reports consume.