W.D. Ross (in "The Right and the Good," 1930) illustrated his qualms with utilitarianism by noting that it would be absurd to keep promises only due to the end result being good (as a strict utilitarian would have to justify it). People don't just decide for themselves, at the last moment, whether keeping the promise is really for the greater good, and then keep or not keep it accordingly. Rather, they keep each promise because they feel a duty to keep promises, in general. (He goes on to "fix" utilitarianism by making "striving for the good" only one of several pressing duties, including "duties of fidelity" such as keeping promises.)
John Rawls (in "Two Concepts of Rules," 1955) defended utilitarianism in this context by presenting promise keeping as a "practice": A combination of "social contract" and "inviolable rule"1. A utilitarian, he said, could choose to subscribe to the practice of promise keeping, as it is a good thing to do. They thus give up any license to change their mind about individual promises, and in so doing, can be relied upon to keep their promises.
To this I say, "Thanks, but no thanks. That works for rule utilitarians, but it leaves us act utilitarians hanging." That's where we get to this document, where I point out that Ross' objection is based on the false notion that we (act utilitarians) undervalue the keeping of promises. I explain that there are several good reasons for us to keep promises, justifiably on purely utilitarian grounds. Ross and Rawls are both dead, so I don't suppose they'll be offended by the style of my rebuttal there.
We all know that keeping promises is a good thing to do. People like Ross feel, perhaps correctly, that "we all know" does not constitute a valid justification under act utilitarianism. They therefore assume that a perfect act utilitarian (if there is such a thing), in the absence of this reason to keep promises, will fail to keep promises, or at least to reliably do so.
I have a special rant saved for people who think that they can point out the flaw in someone else's philosophy by deducing the obviously erroneous actions that would follow from this obviously erroneous philosophy. Reading it will make them weep tears of blood, and they will race to prostrate themselves and beg for my forgiveness before their hearts shrivel into little black raisins and their foreheads explode. I will write it someday.
Certainly, an act utilitarian can come up with reasons2 to keep promises, of sufficient weight that they can be trusted to do so, at least as much as the next person can be. For example:
Given these, it is reasonable for even an act utilitarian to follow a guideline (or maxim) stating that keeping one's promises is a very good thing to do. This leaves us free to evaluate borderline cases according to our own best judgment, without the undue concern that we will take our promises too lightly.
1. Of course, a "rule" in this context has built-in exceptions, so that you can break a promise in extreme circumstances without breaking the rule. What exactly the exceptions are, is one of those questions that threatens to make rule utilitarianism devolve into act utilitarianism if pursued far enough. In my opinion.
2. A competent philosopher can justify, with their philosophy, whatever they feel like doing or thinking. It's just easier for us to do, because we're actually right.