The diffusion behavior of ions in solid-state electrolytes is crucial for the overall performance of solid-state batteries. Studies on the atomistic scale indicate that grain boundary (GB) effects and grain as well as particle size significantly impact overall diffusivity. Recent advances have enabled large-scale, long-time, and high-accuracy simulations of complex atomic structures using ab initio-based machine-learning interatomic potentials. Their massive computational demand, however, leaves a significant gap between atomistic and application scale. In simulations on a larger scale, however, the details of the polycrystalline structure are often neglected or considered under strong assumptions. We propose a microscopic finite element model that considers the fully resolved crystalline structure with an atomistically informed parametrization: The potentially anisotropic diffusion coefficients in bulk and GB are obtained on atomistic scale and a novel collapsed interface element is introduced to represent GB effects. Furthermore, we account for geometric defects such as pores. By means of homogenization, an effective model on the mesoscale is derived. The hybrid bulk-interface diffusion model was found to give consistent results when compared with fully-resolved GB representations as well as atomistic bicrystal data and the conductivity measured in experimental setups. To analyze how material and structural features govern the effective diffusion behavior, the affine structure of the problem is exploited to derive a reduced-order model. The availability of a few full simulations enables highly accurate real-time estimates for the effective diffusion coefficients across the parameter space. This enables rapid identification of preferred diffusion paths and a quantitative assessment of how defects, such as GBs or pores, contribute to the overall diffusivity. Interactive parameter space exploration facilitates in silico exploration of microstructures and their implications on different scales connecting computational and experimental results. This enables us to suggest design principles for solid electrolytes with a special focus on GBs.