Agree with OP, as well as
Currently champions are so overpowering that there is no point in having anything BUT champions.
champions will not be not meant to fight alone, but can play a support role in an army.
IMHO, the most interesting way to integrate champions into the strategy play would for them to support armies of regular units. But right now, even if you have 2 large armies with champs facing off against each other, within one round of battle there's only really champs standing.
However, as I see it the larger problem is this: currently, strategically speaking, war in FE is a lot like stabbing with a spear. Wherever that sharp point (read: champ stack) lands, you win, irrespective of what units they've built. The only defense is for their sharp point to meet your sharp point. The outcome of that single tactical battle then usually decides the rest of the game, and however well Brad programs the AI, my guess is that it's not going to ever be as good in tactical as a decent human player, both because of tactical combat itself, but also because it won't have built heroes so single-mindedly with the greater strategy in mind. For a number of reasons that people have outlined in great detail (levelling, healing, champs not dying), that single win translates into a win in the game.
One thing I haven't seen discussed though is how losing cities themselves contributes. So here goes:
What's the point in having cities/settlements (at least in 0.77)? Capturing shards. In my mind, that's it, the be-all and end-all. The only additional thing I would mention is that cities allow caravans which allow roads which enable better movement. That's also non-trivial.
So, early-game, to capture shards you need to expand your borders as fast as possible, which means teching up as fast as possible, which means you need population. Non-shard cities can contribute by increasing growth and tech. But once the map is settled, and all the shards are in someone's borders, cities become nothing but a means of keeping shards within your borders. In a game with better balance, they'd be unit-producers, and continued tech research would be useful, but here - as discussed - that's irrelevant: whether they produce units or not, those units are going to die before contributing.
Now, since it's champs and magic that rule the map, there's an additional snowball effect after that single spear-point vs. spear-point tactical battle: since you likely just took a city, and if you're smart you targeted one with a couple of shards, they've just lost 10-20% of their shards. Their champs' magic is now 10-20% weaker (less damage, less mana, etc..), and yours is commensurately stronger. Therefore, after that key tactical battle, their champ stack will never win against you again.
So, I would recommend that Stardock look at balancing champs and units better in an overall sense. Right now, as Taltamir notes, the RPG and strategy elements are too far apart. My early-mid game has two separate parts: champ stack running around levelling up, while my pioneers and city building happen somewhere else. Any units I take with my champs essentially become champ-like in that they gain so many HP they'll never die, but any I build and don't take adventuring are there to keep monsters off my outposts and that's it. There should be a reason that champs need to bring a real army with them, and there should be a compelling reason for some champs to spend time at home.
However, balance aside, for the current style of play, the one thing the AI needs to do better is to make sure they win that single tactical battle. That means planning everything from the start with that in mind, and using 'escape' to get out of any battle they're not at least 60% sure they'll win before it even starts.