The new AI risk prediction module: who's testing it and is it worth the uplift?

Anyone running the new AI risk prediction module in ServiceNow GRC? Got the pitch from our account team about "proactive risk intelligence" but the uplift cost is significant.

Need real data points before I push for budget.
* What's the actual input? Is it just scraping existing audit findings and policy documents?
* How are you validating the predictions? False positive rate?
* Integration headaches with existing risk registers and workflows?

If it's just a fancy report generator, I'll stick with our current scripts. But if it actually predicts regulatory exposure from upcoming code changes or deployment patterns, that's a different story. Who's got hands-on experience?