When working on a Microsoft Power BI project, testing the semantic model often takes a back seat, whether due to a lack of time or the effort involved in validating it manually. We often assume that everything should work, but that doesn’t guarantee data quality, referential integrity, or the expected results in the metrics.

In this session, we will explore how to automate tests on semantic models in a reusable way to validate the model systematically. By leveraging the capabilities of Semantic Link in Microsoft Fabric, we will learn to detect issues such as duplicate keys and inconsistent relationships, and, using AI tools, identify discrepancies between the model’s metrics and its data source.

A hands-on session aimed at introducing testing into semantic model development through automation, paving the way for its integration into CI/CD workflows.