The bulk of the AI work is going to be in v1.1. It's almost impossible to make it produce a cohesive strategy given how easy it is for the human player to summon tons of units. It's an O(N^3) issue.
Basically, it boils down to the need for a game AI to be able to have N strategies. In a typical game, say GalCiv II or Civ IV, you'd have the AI have several different general strategies (rush, boom, turtle, etc.). But in Elemental, there are a number of unique challenges.
For instance, you have sovereigns who are very powerful. This isn't a new concept, Total Annihilation had commanders. However, when the enemy commander died, they'd destroy everything around it. Thus, you wouldn't tend to "Com rush".
Another challenge has been the issue of summoned creatures. With the way imbuing works, you can easily crank out a lot of summoned creatures without any real economic impact.
You take these two issues and it makes it basically impossible to write a challenging AI. And yet, asking the player not to make use of the most obviously effective strategy would be insane and "gimping" the game is not an option.
What we are going to do in v1.1 is this:
Your sovereign is the powerful Channeler who generates +1 mana per turn. Each time you capture a shard (which in v1.1 willl have 1 seeded near where you start) it will provide +1 mana also as well as amplify certain spells. When you imbue a champion, it costs -1 mana to maintain. When you summon a creature, it will cost -1 mana to maintain. Thus, if you want an army of spell casters, it comes at the cost of having your sovereign being as powerful as he could because he's sharing his power with so many minions. If you keep it all for yourself, you can only be in one place at a time. It also makes controlling shards strategically meaningful because they are the source of the Channeler's power.
Issues I see potentially.
Weak summons, which are already bad due to arriving as late as they do (require a book of summoning, which is a fairly late magic tech), will now become even weaker. Maybe certain summons should be free, and certain summon shsould cost 2.
Is the limit of one summon of the same type per caster still in effect? I'm assuming yes- and that should stay.
enchants should also cost 1. High essence for the channneler should also raise mana gained per turn, maybe channelers get 1 mana for every 8 essence?
Essence should be the cap in tactical battles, and would tactical spells affect the global mana pool? (I'd almost consider an argument that they should't, but strategic spells shouldn't be castable in tactical battles period)
You are going to have to do something about the # of summoned units if you're going to increase the cost of having a bunch of normal units, which also should be done.
Some have mentioned using a higher system then 1, then balancing based on that. That may be an inelegant but more balanceed solution. Maybe a 10 base, so instead of 1 it costs 10, some summons/enchants cost less then or more then 10?
That said, it is a good thing that we can talk about gameplay changes instead of mostly bug fixes- that's a sign that the issues with the engine- a lot of hard work has been very effective in fixing most of them. Step 1 in making Elemental what it should be, is pretty much complete (there's always bugs to squash, but I think the major work is done now.)