When someone in our cast is voicing a non-human character, we supply them with a few things in the recording booth:
Character bio/description
High level info about their species
If available, we show them some reference art to give them a visual representation of who they are voicing
If they've voiced this character in the past (or if others have done so) we'll play them voice reference to get them into character and vocal range
As for audio postprocessing, that's the domain of the Audio Team who are the real experts here. But to my knowledge, some amount of voice processing is applied to all VO lines (normalization for levels) but many non-human species get additional processes applied (e.g. pitch shift, gutturals added, distortion, echo, etc.) depending on what sort of voice print Audio envisions for that species. And on top of that, if there are special conditions for a situation (e.g. character is a ghost, echo of the past, possessed, etc.) we can layer another process on top.
Again, I want to stress that I'm not the expert on VO postprocessing--our Audio Designers are. But that's how I understand some of the basics of what they do to voiceover lines. The reason I'm commenting here is that quite often a writer/narrative designer will mark a particular character or lines within a scene for conditional voice processing, so Narrative is involved in the process but only at the front end of identifying lines for story or gameplay purposes. After that, the Audio folks take over and make the magic happen.