Cool evolved stars are the main source of chemical enrichment of the interstellar medium. Understanding their mass loss offers a unique opportunity to study the cycle of matter. We discuss interferometric studies and their comparison to latest state-of-the-art dynamic model atmospheres. They show broad agreement for asymptotic giant branch stars. For red supergiants, however, current models cannot explain observed extensions by far, pointing to missing physical processes in their models, and uncertainties in our general understanding of mass loss. We present ongoing imaging and time-series observations that may provide the strongest constraint and may help to identify missing dynamic processes. VLTI studies will remain the highest spatial resolution observations at ESO into the ELT era, complemented by ALMA observations. We discuss crucial improvements in both instrumental and operational areas for surface imaging of cool evolved stars in the era of the ELT.