This question seems fairly opinion based, but I think we can answer this with a basis in psychology.
Have a Cookie!
You did the right thing by rewarding your players for clever play and non-traditional solutions to problems, especially if this kind of out-of-the-box solution is the kind of thing you want the players to do more of. We have to remember that games are teaching tools- every action the players take gets feedback either from you the GM or from the mechanics of the game system itself.
If Krathor the Brave amps up his Strength score, he receives the "cookie" of seeing that he hits bad guys more often and hits them harder. Therefore, he is incentivized to continue increasing his Strength score.
If Slinky the Rogue continues to get cool magic items off of nobles that he pick-pockets, and he is never punished (meaningfully) by this action, he is incentivized to continue pick-pocketing. He gets the "cookie" of cool loot.
So what you have done here is created an incentive for your players to role play with NPCs, even ones they consider enemies, to consider non-violent solutions to problems, and to look for opportunities to make allies. You gave them a "cookie" in the form of a high level NPC ally. If these are things you want your players to do more of, great! You've done everything right so far.
So on to your actual question: should you have scaled the challenges up harder?
Well that was... disappointing.
Let's take a look at what would have happened had you scaled up the encounters. Presumably, this NPC was built up as some kind of badass. If, after getting this guy on their side, the PCs experienced no marked difference in the difficulty of their combat encounters, then their "cookie" means nothing. It is worse than the standard, it is disappointing. The NPC is failing to meet expectations. Now, instead of having neutral attitudes toward the action of gaining allies, the players have an expectation- "Well, it doesn't matter if we befriend this guy, he won't give us any actual benefit."
Instead, keeping the encounter levels the same showcases that their creative problem solving gave them a tangible reward. Success! They are more likely to look for these kinds of opportunities in the future!
This still leaves us with one problem going forward, however.
It's still a game, and the PCs are still the protagonists
The problem with your specific example arises when the high-powered NPC overstays his welcome. Even if the players don't realize it, taking him away after one dungeon (or one story arc, or whenever it is appropriate to do so at the earliest convenience) makes it so that the PCs shift back into focus. People enjoy overcoming challenges; they don't want everything coming easy to them. So while their NPC friend was great for a little while, extended use of him is going to bore the players out of their skulls- once again creating an incentive to NOT look for allies in the future.
As Neil Slater mentioned, this is especially important in D&D 4e, because combat is such a large part of what the game is focused on mechanically, and the combat sections can take up a lot of time. I would get that NPC out of there as quickly as possible after you have sufficiently demonstrated that he was a valuable asset and that the team was smart for converting him to their cause, however briefly.
So in short, you did everything right. Give the NPC an excuse to back out at some point, and let the characters grow from the experience. The players will remember the great benefit they got from looking for a non-standard solution, the next dungeon will seem all the more harrowing without their crutch (even though it's perfectly balanced), and the players have incentives to continue role playing and looking for allies. Everyone wins!