Credit and debit card fraud, in which a malicious actor utilizes compromised card credentials to make unauthorized purchases is one of the most common forms of online fraud and causes billions of dollars in damage every year.

The losses associated with this type of fraud, however, are not usually passed down to the consumer, as major card networks facilitate the reversal of unauthorized transactions if the consumer disputes it with their issuer - these are usually called “chargebacks”.

Since credit card issuers absorb much of the liability associated with cards, they have a vested interest in implementing the tools and technologies to monitor the vast quantities of card transactions processed each day, and filter the ones that seem suspicious and pose an eventual reversal risk.

Preventing card transaction fraud as an issuer is usually done through a combination of machine learning models that are trained on hundreds of signals related to the transaction (e.g. whether the card is being used physically or online, location and amount of the charge, the historical fraud frequency associated with the merchant in question etc.) and static rulesets that are tweaked to catch baseline fraud patterns (e.g. the card is used to purchase high value goods that are being shipped to a location that is far away from the usually shipping address used for that consumer).

How can Sardine help?

Sardine has developed an ‘issuing risk’-specific machine learning model that is trained on historical card transaction & fraud data to assess the riskiness of a given card purchase, either online or offline.

This model can be combined with Sardine’s proprietary, no-code rules engine to pick up on granular and complex patterns and tuned overtime to increase its detection accuracy to provide an optimal trade-off between stopping bad transactions while allowing good transactions to go through.

​
Next Steps

Contact us to schedule a demo and get access to our Integration Guides and API docs.

Business Verification (KYB)Payment Fraud