FRAUDSTERS ARE TRYING NEW WAYS TO DEFRAUD SYSTEMS ALL THE TIME. THE BEST WAY TO FIGHT THEM IS BY ADAPTING TO THEIR NEW TECHNIQUES AS QUICKLY AS POSSIBLE.

We achieve this by employing three processes that are always at work in our system: fraud identification process, fraud detection process, and fraud elimination process. as described in Figure 1.5


1. Threat identification PROCESS

DeepShield's fraud identification process is responsible for identifying fraudulent activity in the system. The algorithms are designed to figure out if something doesn't quite add up. These algorithms detect the "anomalies" and generate a signature of the fraud. Each fraud technique has a different signature and there are infinite possibilities of fraud signatures. 

The fraud signature is used throughout our system and is one of the main building blocks of our fraud protection technology.


2. THREAT DETECTION PROCESS

 The fraud detection process takes the output of the fraud detection process and uses it to conduct a system wide identification of other occurrences of the same fraud. 

If another fraud of the same signature if found, the fraud is logged. Depending on the widespread nature of a particular type of threat, different measures are taken. However, the actions are not the responsibility of the threat detection process.


3. Threat elimination process

Based on the nature and pervasiveness of the threat, the system adapts to new rules through the threat elimination process. 

This is a complex process that works on more than 20 input variables to figure out what is the best methodology to address the current threat. 

This architecture allows the system to be resilient and adaptable to new and unknown fraud threats with minimal to no human intervention.