Metal-oxo species are essential intermediates in a number of bioinorganic reactions, with the most prominent example being photosynthetic water oxidation. Their activity inspires the search for synthetic metal-oxo catalysts for key chemical reactions to address societal challenges in renewable energy storage and sustainable synthesis. For many reactions computational design has become an important tool to identify new catalyst candidates. However, a number of obstacles must be overcome to realize its potential for metal-oxo catalysts: elimination of mechanistic uncertainties, filling of a sparse data space, and proper handling of multiple electronic states. This computational project aims to remove these obstacles through a combination of spectroscopic modeling using multiconfigurational wavefunction calculations, and reactivity modeling using density-functional theory. The multiconfigurational restricted active space method shows excellent accuracy and can be used to assign electronic structure from spectroscopic signals. This modeling will be used to remove mechanistic uncertainty. The project will also establish reliable correlations between reactivity and experimental data, and propose a new route to reveal the role of valence-excited states for reactivity. At a longer perspective, these principles can be used to propose novel metal-oxo complexes with tuned reactivity, designed from both synthetic and fully computational platforms.