This was a good outline of the leadership failures common in large orgs trying to get good value out of data science.

Edit: In summary:

“Any organization’s adoption of AI and machine learning requires three technical tools: usable digital data ..., computational capabilities ..., and a development environment that engineers use to code”