Today’s BI environments have split personalities. They must support the production of routine reports and analyses used every day for decision-making by line-of-business employees, and yet, also enable data scientists and data crunchers to “freewheel” through the data in an unplanned, experimental fashion. What magic is this? How can implementers create a sustainable BI environment with these two seemingly contradictory purposes? Does one replace the other? What are the technological requirements for this new world?