i avoid the "make this work" button like the plague.. they ask for too much and you can do better 99% of the time.
look at strengths and weakness as color coded on teh trade screen... if trading vets look at the MLB column, and if looking to use prospects look a the MiL column for "Red" and "Yellow"
don't just stop with one player of that position.. try a few and read teh feedback to distinguish which the AI likes better... if you don't want to get rid of multiples of one position look for another weakness and rinse/repeat process.
take the players with the best feedback responses and put them together in a trade and see how close you get to a positive feedback from teh AI-GM.
i am not that systematic about it... i typically just start throwing the highest potential of the weaknesses together than start mix matchign to see if they like 1 more than another etc etc... swap out lower potentials to see if they are valued more by chance then switch back, rinse repeat. sometimes having another player or two helps distinguish feedback between 2 players (keep base the same, swap the "1" player out as you go -- the base doesn't have to be part of the trade, just to more easily distinguish value of a 3rd that's swapped out).
always expect to pay more... be willing to walk away if it's too much.. be pleasantly surprised when they don't take enough

it won't happen often even on neutral difficulty.
if your league favors prospects vs vets this affects GM-sliders and how they value players... take a look at the ai gm before entering the trade screen to know what they value.
Scouting innaccuracy opportunities: you can get burned too, but if you spend more on scouting than the AI you shoudl come out on top more often.
there's achance for the AI to overvalue a prospect... which one is it? you can't tell... i guess you could cheat and "act as" and see with their scout who they value -- not sure if that works, but reinforces the idea here. what they see is different and you can take advantage of that in lucky situaitons.
this is way more likely to occur with younger prospects with 1-2 years or less MiL experience.. by the time they get to AA you won't see many that are significantly off from accurate ratings