Episode 3

64% of all AML anomalies can be explained by 5 legit life stories:

• Buying or selling real estate

• Payment for cash-intensive jobs

• Giving or receiving of gifts

• Buying or selling a vehicle

• Construction expenses

​

AML Transaction Monitoring systems look for anomalies in customers’ account that match high-risk typologies. Unfortunately, most of the flagged anomalies are perfectly legit customer activities; best-in-class numbers are 80%, and in most banks the number is somewhere between 95% and 98%. That’s a lot of falsely flagged activities. Is there some magic that can spot those perfectly legit activities before even starting an investigation?

​

Refine’s unique capability to obtain customer explanations to anomalies in their account, done at scale and using a consistent and structured question tree, allows us to create a model with over 2,000 features mapping anomalies to the legit life stories explaining them. The model accurately detects the specific life story behind an alerted activity, and provides evidence to the operations team so they can hibernate the alert. This means we know precisely how many alerts represent each legitimate life story.

​

In fact, just 5 life stories explain 64% of AML alerts. If you’re interested in more stats, sign up to receive notifications for each new Refiner’s Gold episode.