Somatic hypermutation is not that particular wrt mutagenesis itself. The complexity of it shows how evolution cobbled it together. Mutation is selectively increased in highly expressed genes (transcription is literally required) in general, but mutations are not being repaired as often in the immunoglobulin genes. Thus the latter is a refinement layered upon the former, which still had adaptive value on its own. That’s still an oversimplification, because IIRC only one of the two major repair pathways is dialed down.
Exactly. They’ve never been alleged to be random in any other way, except by creationists looking to pummel a straw man.