Artur Tyvaert

Conversational BI is the next big thing. It is advertised as the promising solution to gain new insights from your data. It fills the gap that traditional BI reports fail to discover by directly talking to the end consumer and providing answers to one-off questions. It invites users to explore its capabilities with an easy Q&A format and adapts to changing business requests. Who wouldn’t get excited by all such promising features? However, it is still unclear how the conversational BI architecture should look like. In which scenario’s do we use Databricks Genie, Microsoft Copilot or even a custom build solution? In this session I want to explore the different scenario’s and compare them across different dimensions to answer the question: ‘When do we use which conversational BI setup’? First, I will review the semantic layer in Power BI vs Databricks. How is each semantic layer built? What are metric views and do we need them to harvest the full potential of Genie? What about the development effort and do metric views in yaml have advantages over a semantic layer in TMDL? The next thing that we should talk about is the governance of conversational BI. How do we handle permissions and security? How do user permissions and RLS in Power BI compare with the role-based access control and data masking in unity catalog? The final axes that the end user should consider is the environment. Does Genie have an advantage as it is directly build on the lakehouse? What should we do if we already started building reports in Power BI? Is the conversation history saved in each environment to investigate the needs of end users? If you would like to get answers to all these exciting questions, then this session is something for you! Whether you are new to conversational BI or preparing your first implementation, you will leave with a clear understanding of how each environment shapes reliable and scalable conversational analytics. Requirements: - Have a basic understanding of data modelling Target audience: - Data engineers and Architects - Business analysts who want to have an introduction to the conversational BI architecture and its considerations Main learning objectives: - The different options that are available when you want to do conversational BI - The key decisions that you need to consider when developing your conversational BI architecture - What environment is most suitable for you if you want to do conversational BI