Machine learning techniques, especially deep learning and reinforcement learning, have recently achieved remarkable results in many fields such as computer vision, speech recognition and games. These advances have created huge enthusiasm and expectations in the business world. However, in many business machine learning projects, these expectations are often not met. Based on many projects I led, reviewed or rescued, this article explores the applicability of these advanced machine learning techniques in the business world and how to maximise their results and avoid potential blind spots that they may create.