Because ML (this is what people mean when using the term AI which is ambiguous) needs some data to learn from, and this - apart from the fact that learning on a

Because ML (this is what people mean when using the term AI which is ambiguous) needs some data to learn from, and this - apart from the fact that learning on a living organism like a company can be extremely expensive - poses several problems: 1. What data do you use? Past data of the same company? Data of other companies? Which ones? Past best performers? 2. Actually identifying all pieces of information that are relevant for decision making is a challenge in itself, as we're not talking about internal metrics but also external events and trends. Choosing and collecting all these is a challenge in itself. Moreover, it is not a one-off action, it needs to be reevaluated regularly, i.e. you need to be actively looking for factors influencing the performance of the company. Mind you, not all of them are readily measurable. 3. Part of the work of CEO is creative, i.e. not extrapolating based on past events but looking for completely new avenues and opportunities. Building such a complex system would be extremely expensive, although I agree this challenge is very interesting. Not to mention the fact that an extremely important job of the CEO is to deal with people, not just the data. This point alone can make a difference between a well- and badly-managed company.