Chatting with your data is an increasingly sought-after solution to add on top of our semantic models, and implementing one comes in many shapes. You can build one with plug-and-play components, often in proprietary formats with a lack of flexibility. Alternatively, you could build your own conversational data agent.

In this session, we'll cover multiple approaches to enabling chat-with-your data on top of your semantic models. We skip the simple choice of using Fabric tooling and jump straight to customizing popular external chat agents to be a conversational BI powerhouse, as well as building bespoke applications to give us the exact features we'd want.

Topics:
- Overview of conversational BI tools
- Implementation of ad-hoc analytics and dynamically refreshable reports directly in Claude
- Explaining the internals of using LLM API's
- Building an application from scratch using XMLA endpoint and LLM API's

After attending this session, you'll be well-equipped with evaluating the multitude of options for how to get your colleagues to converse with your data, so that you can make the right choice for your organization and learn the skills to build it out along the way.