PTC Propensity-to-Pay Predictive Model

The Pennsylvania Turnpike Commission (PTC) serves more than 500,000 customers daily, with each one paying tolls through RFID transponders or toll-by-plate billing. Leveraging artificial intelligence, HNTB partnered with PTC to develop a predictive model to understand future cash flows of tolling revenue and maximize revenue collection.

HNTB built behavioral personas segmented by geography and demographics, then overlaid three years of historical toll lane usage and invoice payment information, applying an AI-enabled machine-learning algorithm to look for differentiators and predict payment propensity. The innovative propensity-to-pay model developed from this process predicts the likelihood of invoice payment by transforming vehicles into customer personas and focusing on their choices. Using this data, the PTC can make data-informed decisions on collection strategies for at-risk revenue.

The predictive model forecasts on-time toll payments to the 95th percentile, improving the PTC’s planning capabilities for cash flow management and helping the Turnpike develop strategies to improve toll collections.