Hands-On: Preparing PowerBI Models for Generative BI
An Intensive Hands-On Workshop

 

This course will take place on Thursday December 10th and Friday, December 11th.

Sessions run from 9:00 am – 12:30 pm PT / 12:00 pm - 3:30 pm ET and will be recorded.

 

Whether you call it “Conversational Analytics” or “Generative BI”, data leaders from just about every organization want to enable self-service analytics by combining AI with their existing BI assets.I

For Power BI teams, just because your semantic model delivers accurate dashboards doesn’t mean AI can use it to answer a simple question from a business stakeholder.

Why?

Because the business knowledge and historical context stored in every Power BI professional’s head are rarely available to the AI agent.

In this intensive, hands-on workshop, you’ll learn how to turn an existing Power BI semantic model into a reliable foundation for Generative BI. Using a realistic case study, you’ll work through the practical steps needed to help Copilot understand your business and deliver useful answers.

First, you’ll examine what makes a semantic model AI-ready and why technically correct models can still produce inaccurate or misleading responses. You’ll learn how to enrich your model with clear business terminology, descriptions, and well-defined measures.

Next, you’ll use Power BI’s Prep data for AI features to simplify the AI data schema, create verified answers, and add AI instructions that communicate business context and analysis rules.

Finally, you’ll test your configuration with Copilot in Power BI Desktop using Copilot. You’ll ask realistic stakeholder questions, validate the results, and refine the model and configuration to improve response quality.

You’ll leave with a practical approach for preparing and testing your own semantic models, an understanding of Copilot’s limitations, and a roadmap for extending your skills into Microsoft Fabric IQ and Fabric Data Agents.

You Will Learn:

  • What makes a semantic model AI-ready.
  • Why AI struggles with technically correct models.
  • Enriching models with business-friendly semantics.
  • Simplifying the AI schema.
  • Configuring verified answers and AI instructions.
  • Testing Copilot with realistic business questions.
  • Understanding AI limitations and next steps.

Geared To:

  • Business and data analysts
  • BI and analytics developers and managers
  • Data scientists
  • AI engineers
  • Anyone interested in building real-world Generative BI solutions

Prerequisites:

  • Attendees should be familiar with Power BI Desktop, semantic models, star schemas, and DAX. 
  • Experience developing and publishing Power BI semantic models is recommended
  • No prior experience with Power BI Copilot, Prep data for AI, Microsoft Fabric IQ, or Fabric Data Agents is required.

Laptop Setup:

  • Attendees must have access to a laptop with a current version of Power BI Desktop installed.
  • Attendees must have access to a paid Microsoft Fabric capacity (F2 or higher) and a Fabric Workspace with Admin, Member, or Contributor access assigned to the Fabric capacity.
  • The Fabric capacity must be located in a region that supports Copilot.

     Note on Corporate Laptops:

  • If your organization restricts software installation, Microsoft Fabric access, or Copilot capabilities, contact your IT or Fabric administrator early to ensure your environment is ready for the hands-on labs.

Laptop preparation Instructions will be available on the seminar site prior to the first session. 
There is no time allotted in class for laptop preparation. Laptop preparation is required BEFORE the first session.