The energy distribution grid is a critical infrastructure challenged with shifting requirements induced by the skyrocketing importance of green energy. Particularly, legacy assets-such as mediumvoltage switchgear cabinets and circuit breakersneed to be maintained to prevent energy outages and reduce resource consumption. While related research has abundantly presented algorithms for conditionbased maintenance, no design knowledge is available to prescribe how an information system for this purpose ought to be designed. In a design science research project, we develop an information system for condition-based maintenance of legacy assets in the medium voltage distribution grid that utilizes geospatial data. Our design integrates Enterprise Resource Planning (ERP) functionality with Geographic Information Systems (GIS) and a Machine Learning System (MLS) for predicting outages. We demonstrate a current proof-of-concept and conclude by presenting a set of theoretical hypotheses that can guide the evaluation once the system is available.