Alloy discovery and development is slowed by trial and error methods used to identify beneficial alloying elements. This fact has led to suggestions that integrating quantum theory and modeling with traditional experimental approaches might accelerate the pace of alloy discovery. We report here on one such effort, using advances in first principles computation along with an evolving theory that allows for the partitioning of charge density into chemically meaningful structures, alloying elements that improve the adhesive properties of interfaces common to high strength steels have been identified.