Creating a full-fledged AI is both very cool and, on the other hand, perhaps not really needed by anyone. In any case, none of the chess players are eager to play against Alpha. And chess enthusiasts are still much more interested in watching matches between people, not Alphas.
Therefore, perhaps it is still more promising to simply use new AI cheats (and treat the AI player as nothing more than a "running monster") rather than trying to present the AI player as a real player and a full-fledged competitor.
Added 4 minutes ago
According to my analysis, there is no need for 500 archers. A moderate push in the right direction is enough for the AI to be a formidable force.
You want to present it as a full-fledged player. Maybe someone shares your desire. Then it is difficult to disagree with what you said.
But in order for the AI to simply be a strong "running monster" that can easily defeat a player, advanced algorithms are not necessary.
Whatever the players say, I have a fairly high opinion of how NWC made the AI. It's not that bad.
Added 3 minutes ago
No, I'm not talking about that. Not about the neutrals on the map. I just used the figurative expression "running monster". After all, the AI algorithms for neutrals and the AI player are not different.
Yeah, so you tested this? Did you give the AI cheats and see how the difficulty of the game against it changes?