It’s Time to Re-think AML Strategies

For the purposes of AML, it’s a lot more useful to know who is transacting than the amount that’s being transacted, yet most AML system alerts are built around the latter.Combining internal, publicly available and transactional data, you can complete a full picture for investigators to make a decision on whether something truly looks suspicious.

If you’ve been following the news recently, you’ve probably seen the deluge of headlines surrounding financial fraud and money laundering. Whether it’s Newsweek’s parent company and Danske Bank being charged or Canada, United Arab Emirates and Turkey cracking down on suspicious activity, it’s clear traditional anti-money laundering (AML) strategies aren’t cutting it and there’s a want as well as pressure to do better.

AML systems have traditionally been built piece by piece, essentially copying one approach for different lines of business (retail, corporate, correspondent, trade, markets, etc). This is done after the fact and responds only to a very limited set of behaviors relating to a single incident. This piecemealing complicates the monitoring process by introducing a range of overly simplistic, siloed solutions that can’t help spot new patterns or behaviors, but can only respond to specific triggers. By the time the news stories are circulated detailing the financial crime incident, criminals have already started to change their behaviors to remain under the radar or moved their activity to a new target bank that has adopted similar controls.

Not only is this approach ineffective at recognizing and stopping financial crime threats, it’s also creates a surprising number of false positives – 90 to 95 percent. With such narrow behavior parameters, normal day-to-day activities get flagged as potential crime and analysts are required by law to investigate every alert, no matter how credible. Banks keep throwing more budget and more bodies at the problem, but are still stung by millions of dollars in fines every year and still inadvertently facilitate money laundering transactions equivalent to 2 to 5 percent of global GDP.

What’s missing from current approaches to AML? Context.

For the purposes of AML, it’s a lot more useful to know who is transacting than the amount that’s being transacted, yet most AML system alerts are built around the latter. Anyone can transact in high volume or frequency (both of which would raise red flags) so that information is only useful when combined with outside knowledge about that entity. If they have common connections with people on the terrorist watch list, that’s going to be worth taking a closer look at versus someone interacting with regular business associates.

Combining internal, publicly available and transactional data, you can complete a full picture for investigators to make a decision on whether something looks suspicious, in significantly less time. If this full picture of activity and context is used in the detection and analytical processes, it effectively combines human intelligence with artificial intelligence (AI). This in turn creates context and makes it much easier to detect activity, reduce false positives, and more quickly escalate the number of real incidents that are captured.

Thanks to AI, it is now possible to look for hidden relationships that could reveal nefarious activity and guiding this process by incorporating human intelligence into the analytical streams. AI can pour through more data faster than its human counterparts, making it possible for banks to find out more about criminal networks than ever before. Transactions leave a data footprint, and analytics can start to paint a picture of what criminals are up to – the ultimate goal to be able to map this activity in near real time.

Already, AI augmentation of the traditional AML process has improved efficiency, reducing false positives by over 98 percent and increasing the number of real incidents detected that went completely missed by old systems. This alone has saved millions of dollars in compliance costs and potential fines. With an increased focus on cracking down on financial crime and holding banks and executives accountable for financial crime, institutions can ill afford to stick to the status quo.

The world is a series of relationships, why not model that information? With investments in AI this is becoming possible. Financial institutions can now move from monitoring to being at the heart of disrupting financial crime and the nefarious activities and industries it underpins.