You've spent years building data platforms. Semantic models. Measures with proper business logic. Put them to use in Power BI reports. Now, the latest generation of AI agents can skip the report, query your data and models directly, and not only answer business questions in natural language, but also build visualisations, dashboards and reason over the results.

In this session, you will learn what it takes to build those Conversational BI solutions in your Organization.

We'll look at how semantic models act as the secret sauce for Conversational BI, providing the context, business logic, and guardrails that agents need to produce trustworthy answers.

We'll discuss frontend implementations (SaaS products like Claude, Open-Source frameworks and custom solutions), and confront the strategic dilemma: If the agent is good enough, will anyone still open the Power BI report? Do they even need to?

After the session, you should know:

- How semantic models and their measures, descriptions, and relationships act as built-in context engineering for your agents.

- How to expose your Power BI semantic model to an AI agent using MCP (Model Context Protocol)

- A framework for agent instructions, grounding prompts, and deterministic query libraries built on top of semantic models