Sometimes it is wrong to train a person. You train dogs. It is better to educate people and leave the decision upon them when a particular situation comes instead of programming them to do a particular thing in a particular situation. The innovation is killed in this way. No one will think out of the box and everyone will work as if they were just robots.
Take an example of a football player who is told to cross the ball inside the box every time he reaches the far corner of the pitch. He is trained in that all the time and he might master it after a season or so. However, consider a situation when a defender was prepared for that person to counter that cross. The defender will know what is going to happen so he will be ready to do what he is taught and hence, it will all become monotonous plus robotic.
On the other hand, if the player was told to cross the ball but has an opportunity to dodge the defender/marker and cut inside the box, he will see the scenario and decide about it. This is how the other defender who was waiting for the cross will be puzzled as well. The player might not succeed in dodging the player with his skill but he will at least try to do something different.
Now the situation where training is important: if you were told to print a particular page. You just have to know how to print a page from MS Word and you are done with the task. There is no innovation required because there is no counter reaction. So, you just need to press “Ctrl+P” at the same time and the page will be printed.