Transit Payment Card Risk Scoring for Faster Fare Access
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Open loop transit systems face high default risk due to passengers defaulting on pre-authorized payments, leading to significant financial losses and lengthy queues during peak hours, which conventional risk assessment methods fail to address effectively.
Innovation Solution
A server system uses machine learning models to assess the risk of payment cards in real-time, assigning a risk score and determining a pre-authorization amount, allowing quick access for low-risk transactions while minimizing defaults.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional risk assessment methods are used to prevent default, then default risk is reduced, but queue wait time increases significantly during peak hours
Solution Approach 1:
The system performs risk assessment in advance before the passenger needs to access the transit system. By evaluating card risk profiles and pre-determining authorization amounts beforehand, the system eliminates the need for time-consuming real-time risk assessment during peak hours, thus reducing queue wait times while maintaining reliable default risk management
Solution Approach 2:
The system introduces an intermediary risk assessment layer that operates independently from the main transit access process. This intermediary layer uses machine learning models to evaluate card risk profiles and determine pre-authorization amounts, separating the risk assessment function from the transit access function and enabling parallel processing that reduces overall wait time
2Reliability
If pre-authorization is performed for all passengers, then default risk is managed, but transaction speed decreases during peak hours
Solution Approach 1:
The system applies different pre-authorization strategies to different cardholders based on their individual risk profiles. High-risk cards receive higher pre-authorization amounts or enhanced monitoring, while low-risk cards receive standard treatment. This localized differentiation allows the system to maintain reliable default risk management without applying uniform slow processing to all transactions
Solution Approach 2:
The system dynamically adjusts pre-authorization amounts and risk assessment parameters based on real-time factors such as transaction amount, cardholder history, and current system load. This dynamic adaptation enables the system to optimize transaction speed by reducing assessment intensity for low-risk scenarios while maintaining thorough risk management for high-risk cases
Data Source
AI summary
Methods and systems for managing default risk associated with transit transactions are disclosed. The method performed by server system includes receiving a payment authentication request associated with a transit transaction performed by a payment card and identifying a card status of the payment card, the card status indicating whether the payment card is associated with a risky label or non-risky label. Upon identifying that the payment card is associated with risky label, accessing a pre-auth feature set. Method includes computing, by a pre-auth machine learning model, a pre-auth score based on the pre-auth feature set. Method includes determining a pre-auth amount for a predefined time period based on comparing the pre-auth score with a plurality of predefined pre-auth thresholds and transmitting a risk indication message including at least the card status and the pre-auth amount to the merchant.