Don’t dismiss quite so hastily; it can still be an intermediate form without being a direct ancestor, even if it dates to after a point of divergence. How? Because of punctuated equilibrium, we know sometimes a population changes relatively fast and sometimes hardly seems to change at all. A fossil might represent a later generation of a common ancestor, without having changed much from what we would expect that common ancestor to look like, so in that sense they can be extremely useful to study.
Well, first of all, we do get a few actual ancestor species. But hang on. Who decides what “a perfect fit” is? Based on how evolution works, we certainly wouldn’t expect to see every characteristic halfway between the measurements of the descendants. For example, in our own evolutionary past, our lower bodies evolved for bipedal walking significantly before our brains and diets changed much.
But we don’t just expect to see varying rates of change; we also expect occasional traits that are not passed down to descendant species because they are lost. Does this disqualify a species from being a “perfect fit?” No, of course not!
But how do we tell if traits that don’t seem to ‘fit’ were lost after being present in the direct line, or evolved in a side branch that never left descendants? Well…we just can’t really tell. Unless there are a lot of traits that would have been difficult to ‘lose,’ we just have to guess. So nobody calls anything a direct ancestor most of the time, because it’s so difficult to know for sure, not because we don’t have any good examples!
Additionally, have you ever looked at a really thorough family tree? Most of what you’ll find on it is actually cousins, assuming they haven’t been edited out.