# Prompting for Analytics is Different!

## Метаданные

- **Канал:** Guy in a Cube
- **YouTube:** https://www.youtube.com/watch?v=_uixX5O4jJg
- **Источник:** https://ekstraktznaniy.ru/video/44588

## Транскрипт

### Segment 1 (00:00 - 00:00) []

If you work with PowerBI, you've probably noticed this already. AI is showing up everywhere. People are asking to write facts, explain reports, answer business questions from their data. You ask what seems like a simple question. You get a total that looks reasonable. And a lot of times the numbers are wrong and it's not obvious. It's very subtle. And in analytics, that's the most dangerous kind of wrong numbers. And the answer sounds so confident, but the model misunderstood the semantic model. And once the trust is broken, it doesn't matter how nice your report looks. Ain't nobody going to look at that pretty report. And that's where I'm spending some time going deep on prompting for analytics. If you've tried using AI with PowerBI and something broke, I want to know. Let me know where in the comments below.