Demystifying Fraud Detection

Blog

Demystifying Fraud Detection

Worldwide, fraud is manifested typically after payment credentials are compromised. Having said that, there are scenarios where “friendly frauds” are committed by the card or account holders themselves. Fraud and risk management systems (FRM) the world over attempt to look at electronic payment transactions and sniff out the ones that look like outliers. The primary challenge of FRM systems is how to detect these outliers. Every FRM system is essentially friction in the path of the transaction, so we need a system that can do it without adversely affecting customer experience. Additionally the job of countering real-time payment frauds in even more difficult, as you get only one chance while the data is in motion.

Worldwide, fraud is manifested typically after payment credentials are compromised. Having said that, there are scenarios where “friendly frauds” are committed by the card or account holders themselves. Fraud and risk management systems (FRM) the world over attempt to look at electronic payment transactions and sniff out the ones that look like outliers. The primary challenge of FRM systems is how to detect these outliers. Every FRM system is essentially friction in the path of the transaction, so we need a system that can do it without adversely affecting customer experience. Additionally the job of countering real-time payment frauds in even more difficult, as you get only one chance while the data is in motion.

The most simplistic way is to check the details of a payment source identifier, like card number or account number that has previously been put in a hotlist. If the source identifier exists in the hotlist, the transaction can be stopped immediately.

The next level of risk assessment is when the transaction is assessed in the context of the recent activities associated with the payments from that card or account. For example, in a POS terminal, it is possible that the card is swiped twice due to some error of the clerk or malfunctioning of the magnetic stripe reader. However, it is unlikely that the consumer will keep on providing the PIN in multiple cases and authorize many transactions from the same merchant within a short span of time.

The other context for assessment of risk is behavioral. If the system tracks the average amount of withdrawal from an account every month and suddenly sees a transaction way too deviated, it triggers an alarm. Similarly, if a card is normally used in an ATM within an area, then sudden usage at a distant location looks suspicious. Here the system tracks the behavior of the card or account holder and compares to see the deviation from normal “behavior”. However, a first time user will not have an activity or behavioral profile, so we need to treat such use cases separately.

If you are the bank that holds the relationship with the card or account holder, you are privy to KYC information and the individual’s long term transaction history. Hence information about the individual can well be used to get a better sense of the spending patterns and hence detecting anomalies. This is not the case for “hubs” that sit in between the merchant and the bank of the consumer. However, there are certain methods that can help relate different payment sources looking at the payment messages. This can then help build an individual persona view out of these payment sources, which in turn can then be used to identify anomalies with better precision.

RS IntelliEdge™ is built with all these capabilities to detect fraud and suspects using hotlists, activities, behavioral and individual based filters for both real-time and offline systems across multiple digital channels. The system creates a seamless experience with an unnoticeable response time of a few milliseconds.

Leave a Reply

Your email address will not be published. Required fields are marked *

Comment moderation is enabled. Your comment may take some time to appear.