Smart data collection and geospatial analysis have the potential to allow providers to optimize the distribution of agent networks. This report notes that in the areas studied 52% of the rural population lives more than five kilometers from their nearest agent or bank branch. Through geospatial-modeling exercises, we identified 52 alternate service delivery points that, if acting as agents, could bring all the villages within a two-kilometer radius of a financial service point. The report also discusses agent segmentation to ensure agent viability and reduce agent churn. Introducing segmentation in the geospatial model resulted in the optimization of the agent network with fewer agents required to serve the target population.