Many cellular functions involve the assembly of biomolecular complexes, a process mediated by water that gets displaced as subunits bind. This process affects water frustration, that is, the number of unmet hydrogen-bonding opportunities at the protein-water interface. By searching for least-frustrated aqueous interfaces, this study delineates the role of frustration in steering molecular assemblage. The search entails a trajectory sampling using a functional that measures the gradient of frustration and computing the resulting non-Debye electrostatics within relaxation times for coupled protein-water systems. The minimal frustration principle is validated against spectroscopic measurements of frustration-dependent dielectric relaxation, affinity scanning of protein-protein interfaces, and NMR-inferred association propensities of protein-complex intermediates. The methods are applied to drug design, revealing the targetable nature of the aqueous interface.