Planning for artificial intelligence should start with whys, that is, the use cases. After that, think about the data strategy to train algorithms and support ongoing AI deployment. Without a solid data strategy, an AI project could fail because of erroneous output.