Mastering Semantic Modeling in... Azure Databricks?!
About the Session
Semantic models are old news in the BI world — but the models built for dashboards weren't designed for AI Agents. You can't manually model every relationship, yet every company still needs to deliver trusted, governed information to every user. In this session, I'll cover the real challenges of semantic modeling, the best practices that hold up at scale, how your model shapes the Genie One ontology, and how to serve it wherever your users already work — Excel, Teams, or virtually anywhere.
... YES! It also works with Excel.
Talking points:
- The problem of inconsistent metric definitions
- How to give the business context to an Agent
- How to make teams agree on a definition, or at least start talking about it
- Keep the numbers in BI consistent with Agents
- Open source initiatives and OSSIE
- Last, but not least - get data on Excel
About the Speaker
I'm a Solutions Architect at Databricks, based in Berlin. I've worked across Brazil and Germany with multiple Fortune 500 companies — from small BI teams to platforms serving tens of thousands of users. My thing is data warehousing and semantic modeling: the metrics, relationships, and business logic that turn raw data into a layer people actually trust, across Power BI, Agents, and the Lakehouse. Get the data model right, and you save thousands of expressions and one-off fixes down the line. Mostly, I love helping teams modernize their BI and data warehouses without losing the meaning their business runs on.