If they have to verify the results every time, what is the point?

have some quarterly stupid “trainings”

Feeling this in my bones, executive just sent out a plan for 'fixing' the fact that the AI tools they are paying for us to use are getting roasted for sucking, they are giving the vendor more money to provide 200 hours of mandatory training for us to take. That's more training than they have required for anything before, and using LLM tools isn't exactly a difficulty problem.