Modern quantum chemistry reached a remarkable level of description of atoms and molecules and their interactions. Theoretical approaches are particularly helpful when experimental studies are hampered or slowed down due to a trial-and-error approach. In such cases, computational chemistry can provide the much sought-after understanding of molecular properties and reactivity. Unfortunately, conventional wave function models are too expensive for large-scale modeling or require user control on an expert level, while density functional theory may predict unreliable properties. To break the current paradigm of computational chemistry, novel and neat approximations are desirable. One such innovative approach models many-electron systems using electron pair states. Current electron-pair methods are, however, insufficient to reach chemical or spectroscopic accuracy for large molecules of organic electronics and must be extended to (i) accurately describe electron correlations beyond the simple electron-pairing effects, especially in cases where conventional corrections break, (ii) reliably predict molecular properties of both ground and electronically excited states of closed- and open-shell compounds, and (iii) provide an intuitive and black-box platform for non-expert users. These goals will be achieved by (a) dressing electron-pair states with information extracted from multi-reference wave functions using a bottom-up approach, where each step systematically improves the accuracy of the previous model along the ladder of approximation, (b) designing a black-box interface to automatized quantum chemistry calculations using concepts of quantum information theory, and (c) elucidating the structure-properties relationship using the picture of interacting orbitals. The synergy between an inexpensive but reliable quantitative description and the qualitative interpretation of molecular interactions will accelerate the discovery of new materials in organic electronics.