How do you define success metrics for AI projects that have probabilistic outcomes?
No STAR scenario available for this question yet.