Although machine learning can have a significant positive impact on organizations, nearly half of all machine learning projects fail at most companies (even those with data-driven cultures). In this post, Barr Moses (Monte Carlo) and Manu Mukerji (8×8) explore why. They discuss the misalignment of business objectives and technical approaches that may occur and address technical challenges like model generalization and deployment. They then present a few tactics that can increase the likelihood of success, such as leveraging the cloud, using CI/CD, and implementing tools for observability and monitoring.