ATM providers were struggling with impaired profitability (replenishment cost, finance cost, and lost revenue) after cash withdrawal patterns changed post a systemic event ( Demonetization of a high-value currency note). Its semi-manual process for forecasting and replenishment model was unable to rise to the challenge post demonetization.

The project objective was to optimize the cost of servicing (Replenishment Cost, Finance/ Inventory Cost net of lost revenue opportunity). The first task was to produce more forecast of expected cash withdrawal at the ATM/ Day level. Then using this forecast along with other information like cash present in ATM, cash supply reliability, operational reliability, replenishment cost structure optimize the daily fulfilment schedule.

The framework consisted of a demand forecast model and then replenishment optimization based on constrains. Same was implemented to work on daily planning cycle.

Overall 18% reduction in ATM servicing cost.