What do you mean by purpose? That is a squish word.
I’m rather saying that the first individuals to claim the selective benefit of a beneficial mutation are more likely to be more evolvable than their neighbors. So selection for new/improved function intrinsically selects for evolvability. That is quite a list, because all those things make it easier to evolve a beneficial mutation, and more likely therefore to end up selected.
I would agree that the much of this process might be neutral initially. However, it causes the beneficial mutations that contribute its selection. So we can’t quite call it neutral draft. There is a causative relationship between evolvability and selection.