When an AI system makes a consequential decision, someone in your organization needs to own the outcome. Is that clear?