I like what DPsycho is saying, but don't wholly agree. (In realistic terms, )Just because you realise that it would be a good idea to build a wall, doesn't mean your Wizard is going to be able to do so all of a sudden.
Way I personally see it would be:
Brick-matching: Your scouts identify the enemy structures, possibly finding weak points and thus reducing the effective strength of the structure (-enemy wall/tower).
Gems-matching: You spot the sorts of magic the enemy has prepared for you and um, the next thief you send out grabs a little more swag than he normally would..? Not sure about this one.
Recruit-matching: You've spotted the enemy forces, seen what you're up against and thus are better able to defend yourself (+wall).
Mixed: Mixed cards are too varied to determine what exactly would've happened in a realistic situation, so the stock increase seems fine by me. Perhaps this could somehow reflect how this kno... aw, to hell with this. It just seems sensible :P
Zero-cost: Your scouts failed to spot anything useful.
EDIT: I haven't really thought through how these effects would affect gameplay. And I also have nothing against the current system, except that the brick- and recruit effects are quite similar and could provide more variety.
EDIT: My suggested gems effect is a remark on the potential absurdity of adding tokens, not a fully serious suggestion. I couldn't think of any better, mind you.