I cover so many different regions that it's not really practical to use a key in every instance (that's if they are available at all), what I'm typically doing is a combination of radar searching to generate a list of recorded species (by herbarium specimen) then narrowing it down by family (or genus If I can). Once there I can then consider a key, or more often I'll consult descriptions and knock off species that don't show the features I'm observing. For this case, Senna artemisioides was my guess from the onset so I once I confirmed it was the species in the area I went and determined the available subsp of which this is the only one with those lovely broad leaves.