Bear in mind that DA simply does not have a sophisticated AI opponent like TOAW (another operational level wargame) does. In that game, each unit in the scneario can have multiple objectives which the computer will fight to seize and hold. It is also smart enough to maneuver its forces in order to re-capture a location that has been lost, no matter where it is. Furthermore, there are multiple objective "tracks" available to the scenario deisigner. The designer can set certain conditions which, if met, will trigger the event engine to switch the AI to a completely separate AI track. In each one of these tracks, each individual unit can have separate orders. This can get really complicated to set up for larger scenarios, but if affords a motivated scenario designer the ability to craft a formidable opponent that can dynamically react to major developments on the battlefield.
DA's artificial opponent is really simple. All it is really capable of doing is moving more or less in a straight line from the first designated loaction to the next. And it can only have a total of two objectives per unit. If every single one of these is subsequently seized by the enemy, the AI will sit on its final objective and make no attempt to re-capture them. In effect, DA's AI does not fight for objectives at all, it simply moves from waypoint 1 to waypoint 2 by the shortest possible route. Once you understand this and have played against the AI a number of times it quickly becomes apparent that the AI is good for learning the game and not much else. It is 100% predictable every single time and cannot dynamically react to any enemy development or strategy.
DA is, however, a fine game when played against another person. In my opinion, DA doesn't get enough respect for what it is able to offer in PBEM mode. It's a fun game this way and more people should try it.