These are good posts and I want to add some of my own issues too:
#1 N^3 game mechanics. These destroy any hope of a good AI. A game having ONE of these would wreck the AI, Elemental has multiple of these. Specifically:
a. Enchantment slots that are PER caster.
b. Individualized mana pools PER caster.
c. The requirement that one must control map resources and those resources combined with the ease of building cities which requires the AI to spread its forces making it vulnerable to force projection (AIs are awful against defending against force projection -- that's the real reason why game designers hate "stacks of doom".
d. Multiple technology paths that result in game winning results with no counter enabling easy force projection.
e. No counter-spells, making the AI vulnerable to force projection even more.
Anyone one of these 5 things alone would make it very hard to have a good AI. All 5 need to be addressed.
If anyone is tempted to ask "Well, you designed the game! How come you didn't see these?" you now know why most games are sequels. It's hard to design a new game (and MOM has multiple N^3 mechanics too which is why its AI would never be competitive). But all 5 of these things we are going to address in Elemental v1.x.
#2 Lack of player control over pacing. It's all about resource getting rather than city strategizing. Elemental needs to hand control to the player.
#3 Lack of a means to move around the map quickly late game without cheese.
#4 Tactical Battles are not fun for me. I haven't found any game yet that I like tactical battles in. I've turned over the design of tactical battles to others on our team because my distaste for them really showed up in v1.0.
#5 Trigger system lacks sufficient conditions. This means I can't easily create random events to spice things up.
All 5 of these things are required to be addressed.