I have to be honest I really like the predictability aspect of having no AI.
The thing I was going for at first was an area selection before the shop. Some areas would have more mega units, some would be better suited for close combat, some would feature something interesting in the level like a hole in the middle, a lot of walls, special enemies with very strong effects that makes you rethink the whole floor....
The predictability comes from the fact that I suck at chess. I want a game where I and only me get to do the plans in order to win a fight. If I lose, it is my fault and not some RNG. It is not well executed at the moment, but I feel that AI would not go into that favor.
I'm thinking about how to make the more interesting and have multiple directions I could go for, I'm starting to like the idea of getting the game developped further but am very scared of the scope (and a little bit of impostor syndrome + perfectionnism which would just freeze my ability to publish anything that isn't perfect). I like the game, it has issues, technical ones and some more design heavy. If I come across some clean solutions it could really become a game on it's own and not some kind of clone!