One way you might propose to answer the question "What should I do?" is to say "Promote what matters." I've already said that sometimes philosophers seem to think that promoting what matters and doing what one ought are the same thing, which inclines them toward (however subtly) consequentialism. A consequentialist believes that we should find out what matters, what kinds of states of affairs are most valuable, and go about promoting them. Utilitarianism is a common type of consequentialism, which is the view the only thing we should promote is the amount of utility (often conceived of as happiness, or pleasure) in the world. There are (to my mind, more plausible) versions of consequentialism that say that lots of different things matter, like happiness, success, relationships, art, etc. These values might not all be commensurable, such that a loss in one is adequately made up for by a gain in another, but ultimately what we should do is promote these values as much as possible.
Non-consequentialists do not see moral obligations in terms of this kind of value maximization. So one question that naturally arises is, what other ways can you answer the question of what one ought to do?
T. M. Scanlon (most particularly in What We Owe to Each Other) is a good example of a philosopher who has taken this question head-on. Korsgaard also discusses it at length, and in a quite compelling way, but many aspects of her answer are quite involved and rely on other controversial arguments. When addressing a question like this, its often best to just stick to relevantly common premises.
One value that Scanlon proposes that is best responded to in a non-maximizing way is friendship. What this means is that to properly value friendship, one cannot simply go about trying to make sure there is the most amount of friendship in the world. Suppose I think friendship is really valuable, so I start a website dedicated to people making friends, and I start a campaign for recognizing the importance of friendship in our lives, and try and make it as easy as possible for people to be able to spend time with and keep obligations with their friends. But all the while that I'm doing this, I don't have any friends of my own. Most people would think I've made a mistake. I got something right, in that I recognize that friendship is important, and I've done a lot to help others. But it seems I haven't really understood the value of friendship unless I've tried to have friends.
I think that captures the intuition pretty well, but a decision scenario might make it clearer. I believe Scanlon gives an example similar to this. Suppose I have a friend, and I've made a promise to this friend, and the continuation of this friendship rests on the keeping of this promise. Imagine also that if I do keep this promise, through no fault of my own, and no wrong-doing on my or my friend's part, two other friendships will end. These are friendships of people I don't know, and have no special obligations towards. Now, regardless of what someone thinks I might actually ought to do in this situation (though, the facts as the are, I think I most likely ought to keep the promise) we can ask the question, what would I do if I were most properly valuing friendship? It seems to me that there is something intrinsic to the value of friendship that demands being a good friend rather than promoting as many friendships as possible.