This reminds me a bit of dating advice given by an attractive person, except it is work performance advice given by a (probably) unusually intelligent and socially capable person.

It is true that most corporations are riddled with less-competent individuals (around half below average, strange to say). But that is expected and generally doesn't matter; corporations are built on the assumption their employees and managers are average. The competition is also made up of average workers and managers [0]. But the junior and senior managers are embedded in a web of relationships and connections and that is generally their value proposition to the company.

So the riddle is not identifying that nobody knows what they are doing, which is easy and obvious, but quickly identifying what can be done technically and then selling it in a way that befuddled management simultaneously agree with and don't feel threatened by. It is not easy, it requires treading a very precise thread of social decisions. It is such a rare combination that when someone turns up who can do that they are justifiably going to be paid absurd amounts of money.

Society is built by a small number of serial overachievers. I expect they do spend most of their time confused why everyone else keeps making obvious mistakes and getting predictably bad outcomes.

[0] Breaking this assumption is where I expect the big gains from AI to come from in the corporate world.