One difficulty is that resonance, aromaticity and some more are additions to older theories that imagine bonds between atom pairs - which are very useful theories - while QM and orbitals tell that electrons spread over more atoms. See: molecular orbitals.
Though, I haven't heard about a manageable version of molecular orbitals that would make quick, qualitative and reasonably reliable predictions about reaction mechanisms. Is there any? Or is someone willing to start this looong development?
Until then, bonds between atom pairs, plus resonances etc, which are known to be wrong, must be used because they bring results.