Data science, broadly defined, has been around for a long time. But the failure rates of big data projects in general and AI projects in particular remain disturbingly high. And despite the hype (e.g., “data is the new oil”), companies have yet to cite the contributions of data science to their bottom lines. What is going on?