ArgoNeuT has reported the first fully-automated reconstruction and selection of GeV-scale neutrinos using data from the NuMI beam line, yielding a measurement of the $\nu_e + \overline{\nu}_e$ scattering cross section on argon. Electromagnetic shower classification tools were developed for identifying GeV-scale $\nu_e$-like interactions among complex backgrounds in LArTPCs. These techniques expand upon traditional $\nu_e$ classification methods that rely on topological and calorimetric information local to the neutrino interaction vertex. We will discuss how these results inform calorimetry-based $\nu_e$ selection in current and future LArTPCs and strategies for further development.