Cash Flow Optimization in ATM Network System

The organization aimed to reduce operating costs associated with provisioning cash machines.

Our ML model used the actual daily data of ATM cash withdrawals for further data analysis:

• data evaluation; setting the requirements and success criteria; data loading, depersonalization, and data enrichment; experiment procedure agreements.
• segmenting the research objects; training, testing, and evaluating the quality of the model.
• automated data loading or model deployment in the customer’s environments; regular quality control by A/B testing.
• technical support of the model and optimization of new data entry.

The implementation resulted in automated cash demand forecasting with an accuracy level over 96%. It achieved significant operational cost reductions, including:

• Lowering the allocation of funds by up to 30%.
• Reducing cashback by up to 40%.
• Minimizing out-of-cash downtime to as low as 0.2%.