How does machine learning assess payment risk?

The latest version of Adyen’s risk engine uses two machine learning models designed to mitigate fraud while keeping conversion high. 

The fraud risk machine learning risk rule evaluates the fraud risk of transactions and classifies them with a distinct risk level. The evaluated risk level indicates the likelihood that a payment will result in a fraudulent dispute. When the risk level is above your risk threshold, the transaction is blocked before it’s sent to authorization.

Currently, the fraud risk machine learning model evaluates credit/debit card and PayPal transactions that could result in a chargeback. If you process other local payment methods, we recommend adding custom rules to complement your risk profile.

The bot attack machine learning risk rule detects and blocks transactions that come in unusually fast with suspicious payment properties. This can indicate scripted attacks like card testing and bot attacks. The bot attack machine learning model evaluates transactions before they are sent for authorization, saving you on excessive retry fees.