What you seem to be underwhelmed by is what people have been asking for since the beta, even. The main complaint about the current diplomacy system was always that it was one sided, you could never give AI missions.
Yeah, but it should have been in a patch. It's not even worth $1 to me. If I want an AI to attack someone, as long as they're an ally, I just need to click on one of their planets, then say, "Set Ally Attack Target." Sure, I can't specify whether I want ships or civilian structures, but so? A quick hack to do the same thing might be to let you click an enemy planet, then do the same thing, for some monetary cost. Yeah, not as polished, but the infrastructure is already there.
And requesting resources from a computer, while doable, isn't particularly useful because I always have a stronger economy than they do.
I just don't see this affecting gameplay any more than superficially. It might make it so CPU won't give you missions until later in the game (which is bad), and it might allow you to have non-ally players attack a given target. Great.
But not really gamechanging, and even for a team of 11 working on other projects, I fail to see why this has taken so long. Unless the way they had coded Sins from the start made them have to rework a lot of code in order to do this, it really should have been a couple of switch statements, a few UI modifications, and then using existing infrastructure. Also, coding a bit of simple AI to decide whether the comp thought it was worth it. I'm not pulling this out of my ass, I'm a programmer. The only other way I can see this taking so long is if it sat on the backburner for a long time, because I could code something equivalent in C++ in an hour or two, it just wouldn't be part of Sins.
Details haven't really been announced, so things might change. But it's underwhelming any way you look at it. All of the features that have been added (AFAIK) are basically allowing you to do things the computer can do... after you research it.
The horse armor in Oblivion was more game-changing, although I admit the dev. time on that was clearly near zero.