Yep...it's that "thinking" thing that would keep it from being automated.
All the color scientists working on color management would LOVE to be able to come up with something that could adjust an image based on output parameters. But the problem is that you needs eyes and a brain to be able to make the final decision.
You can already create an action that will auto-create a pair of adjustments (curves and Hue&Sat) and make a layer group and even pre-select the correct soft proof condition after duplicating the original image. Heck you could even turn the action into an F key. So then all ya gotta do is hit a key then activate the adjustment and twiddle.
But to have some algorithm make the adjustments would mean teaching it image aesthetics which is a bit harder than making profiles.