Machine Learning Explainability for External Stakeholders