Introducing Training Day 1 at Power BI Next Step Conference 2026. Over the next three weeks, we’ll be announcing all three training days, and we’re kicking things off with a day that gets straight to work.
You’ll walk in with an ugly, undocumented Power BI report (we’ve all seen one 👀) and walk out knowing exactly how to use AI tools to fix it, faster and smarter.
This isn’t a demo session. This is a real project with real legacy issues: logical errors, misleading comments, performance problems, the whole mess. Your job? Discover, clean up, and improve using the LLM of your choice.
Led by Eugene Meidinger – BI consultant, Pluralsight author, and 10-year data veteran – the day takes you from zero AI experience to a confident, practical framework for deciding when to trust your LLM and when not to.
You’ll leave with a clear rubric for evaluating any AI task:
- ✅ Ease of Verification
- ✅ Safety
- ✅ ROI
The labs cover everything from LLM theory and legacy report discovery to model cleanup and advanced use cases. Built around GitHub Copilot + VS Code, but any IDE and LLM agent will work.