Less than successful machine learning (ML) or artificial intelligence (AI) projects fall short because they fail to incorporate domain knowledge, suffer from poor data quality, deliver an inadequate business case or struggle with models lacking robustness.