Visitor Forecaster Qorus-Infosys Finacle Banking Innovation Awards 2024

The core concept of the Visitor Forecaster project is to forecast the number of visitors for each branch of the bank for every day of the upcoming week. These forecasts are further broken down into four equal time periods per day. The main objective is to assist the Capacity Planning team in determining the optimal number of tellers required to be scheduled in each time slot ensuring efficient staffing.

The reason why this project was initiated was to address the challenges related to unpredictable visitor traffic at branches. Fluctuations in the number of customers visiting branches often resulted in either understaffing, causing long waiting durations and poor customer experiences, or overstaffing causing unnecessary operational costs. By successfully forecasting the number of visitors, we aimed to optimize staffing operations, decrease operational costs and improve customer experience.

The development and implementation of the Visitor Forecaster involved collaboration of multiple departments. The Customer and Service Analytics was in charge of problem definition, data preparation and model development of the predictive model, while IT ensured the integration of the model into systems. The Capacity Planning team provided us their expectation of the model outputs and how they'd use them.

The Visitor Forecaster project has been succesfully completed and is currently operational. The Capacity Planning unit is actively using the forecasts to schedule the tellers, resulting in optimized staffing levels and improved customer service. The project seems to meet its objectives and may be the groundwork for future enhancements and applications of machine learning in other ares of our bank's operations.