Agree with other answers here. But it's useful for just some of the "easier" use-cases. Just throw an example here. Sometimes when proving the upper-bound of a value (runtime, probability of error, expected runtime, etc) it's helpful to use derivatives to show that a function is decreasing/increasing.