IMO, the AI in MTW did a much better job of dealing with human armies with archers than in RTW (on the offense). I have many vivid memories of the AI forming up en masse outside of archer range before charging me. I have yet to see that done effectively by the AI in RTW.

The frustrating thing is that it would be so much more effective in RTW due to the friendly fire issues. In MTW, it was very safe to have a wall of spearmen with your archers *behind* them (until the enemy mass hit your line). In RTW, placing archers behind another unit is a sure way of killing a bunch of your men except in very specific situations. This forces you to put your archers in front of your front line. If the AI did the "form up outside of arrow range then charge" maneuver from MTW, you would get only a couple volleys before being forced to retreat your archers behind your front line. The mass would then hit your front line, and you'd have a bunch of useless archer units for the rest of the battle.

In my experience, RTW AI armies tend to attack much more piecemeal, which allows you to tear them up. They'll charge a few units at your archer line, you pull them back, kill those exposed enemy units, rinse, repeat. Half the time, the lured enemy unit is the enemy General, which makes it even worse for the AI (apparantly the General unit gets so fixated on the archer unit they ignore the mass of spears that the archers just ran behind).

Of course, this is only on defence. On offence, I can pick apart the AI with archers just like in MTW. Actually, I find the AI in RTW tends to send cavalry after your archers quite nicely (one of the few good ways to minimize this tactic). Now if only it would figure out when the archers have reached safety and turn the cavalry around ...