So good news is I'm focusing even more on realistic expressions and responses... bad news is this comes at a cost of additional development time as it absolutely kills the render:dialogue ratio. In EP1 I had this at around 1 render per 3 or 4 lines of dialogue. So far in EP2 I'm closer to 1 render per 1 or 2 lines of dialogue since there are so many times when someone should either change their facial expression, what they're looking at or how they are standing.
As for the girls, the three will obviously be a focus as they're the core of the family, but at the end of the day my main focus is on writing a compelling and interesting story. This means there will be times when the one or even all of the girls aren't around, but this will always make sense. I'm already getting decent at mixing together multiple renders so more characters are on screen than my GPUs can handle in one scene so I don't plan on letting the hardware limit how I tell the story.
As for LIs, there are definitely going to be more added, but I don't plan on going crazy and not all of them will live with the MC or be around constantly. I write my characters as if they had their own desires, dreams and lives, and sometimes those mean they need to leave the nest for a bit. I for one hate when I'm reading an AVN and I think to myself 'Where is so and so? Why aren't they in this scene?' and my goal is to keep that from happening in my own work so we'll see how good of a job I do

If you ever do find yourself asking that question then please let me know so I can do a better job with the next episode.