In 2012, HSBC reached a £1.2 billion settlement with the U.S. government over the bank’s money laundering controls.
The massive figure sent shockwaves through the financial sector. It was clear from the outcome of the case that banks would have to commit significant resources to their anti-money laundering (AML) efforts or face steep penalties from regulators. It was the corner stone event that led to the creation of an army of financial crime fighters.
In the nine years since, we have seen AML teams globally, grow in size and in cost. Transaction monitoring in AML has become a core capability for those teams, but also a major cost center for those businesses.
So, the financial sector began to look for new technologies that could create efficiencies in the way they fight money laundering. Along the way, a real opportunity has emerged to position AML alongside other financial crime-fighting teams to assess and manage risks holistically, we call this FRAML.
In this article, Featurespace Head of Financial Crime Araliya Sammé discusses how AML transaction monitoring fits into the larger picture of financial crime management.
What is transaction monitoring?
Transaction monitoring is when a financial institution looks at the financial movements its customers make — their transfers, their withdrawals, their deposits, everything — to determine whether that activity poses any financial crime risk to the business. This monitoring is traditionally batch-based but can now be done in real-time.
For a bank’s compliance teams, transaction monitoring is a critical capability. AML monitoring lets those banks get a comprehensive picture of customer activity so they can assess an individual customer’s risk, flag suspicious activity for further investigation by competent authorities, and anticipate potential future instances of money laundering.
The systematic monitoring of transactions has undergone major changes in the last few years. Until recently, AML teams relied on rule-based solutions, which can be good at flagging activity but myopic in the way they assess risk. Rule-based solutions also tend to throw up lots of false positives.
That has proved an unsustainable way to fight money laundering. Banks don’t want rules-based AML transaction monitoring tools. They want innovation. That’s why most financial institutions have adopted, or are in the process of adopting, machine learning tools that enable more effective, more efficient responses to money laundering. A recent (June 2021) study conducted by ACAMS, SAS and KPMG surveying more than 850 compliance professionals shows that 57% of respondents have either deployed artificial intelligence or machine learning in production, or intend to in the next 12-18 months.
How does AML transaction monitoring work?
Transaction monitoring works by having software pore over the current and historical information about a customer, as well as their transactions, to paint a full picture of that customer’s activity.
It’s illuminating to see how this capability fits into evolving models of financial crime prevention.
Previously, AML teams would use rule-based solutions to flag money laundering activity. If X transactions happened, the solution would send Y alert. The AML team could be siloed off and spend their days investigating alerts as they came in. For a while, the banks and the regulators were happy with that workflow (though members of the AML teams certainly less so given the amount of time spent on unproductive alerts).
But as with all financial crimes in the last decade, money launderers got more creative, legal and regulatory requirements expanded, and with that, investigating cases became much more complex. The number of false positives grew. Challenger banks emerged with different priorities, different approaches to innovation, and different AML needs than a one-size-fits-all system could provide. Meanwhile, compliance teams at major banks massively expanded their compliance teams to thousands of individuals and created offshore service centres to deal with the flood of alerts. Compliance teams grew from a few dozen people to hundreds of people as major banks ‘threw people at the problem’.
Quickly, financial crime teams found it was too expensive and unwieldy to reactively chase AML alerts. This is when a new risk-based approach to fighting financial crimes began to emerge — and it has changed the way banks approach anti-money laundering transaction monitoring.
How AML transaction monitoring can help your business
Contemporary AML transaction monitoring relies on machine learning to build holistic risk profiles for financial businesses. The Adaptive Behavioral Analytics that our team has built learns from customers’ transactions and behaviors, and it helps financial crime teams recognize the difference between normal customer behaviors and suspicious behaviors.
This is creating all kinds of new opportunities:
- Alerts are being triaged so that the AML teams can respond to high-priority cases first, and create an optimized workflow according to the risks the alerts represent.
- Monitoring is being tailored and adapting to new digital products and services offered by new market entrants or established institutions, rather than trying to press them into a classic retail banking mold.
- Compliance teams can segment the types of money laundering activities so that they can deploy the correct strategy for investigating insurance money laundering, for example, which differs from money laundering done through a retail banking account.
- Fraud management and AML activities are being integrated into FRAML strategies, in which data gets shared across silos to give each team a better perspective of the business risks they’re fighting.
- Explainable models are helping both regulators and bank executives see exactly how those risk assessments come about.
By not merely monitoring individual transactions one at a time but learning from the patterns of behavior that emerge from transactions, banks gain a real advantage over criminals.
HSBC: Insurance-focused transaction monitoring solution powered by machine learning and cloud
Although its current transaction monitoring processes are compliant with its policy and are largely effective, HSBC wanted to strengthen, enhance, and optimize its manual and semi-automated AML transaction monitoring control environment by implementing a cloud based automated solution for their insurance business. They selected Featurespace for several reasons:
- The solution is specific to their use case
- It leverages the power of machine learning to reduce false positives and increase risk coverage, delivering a more effective monitoring solution
- It streamlines the investigation process by prioritizing alerts generated by both AI and rule-based scenarios
- It delivers explainable AI
In 2021, HSBC won Celent’s Model Risk Manager Award for their innovation in transaction monitoring with ARIC Risk Hub.
From transaction monitoring to holistic risk management
By taking a hybrid machine leaning approach to fighting money laundering, banks move AML away from being a cost center and toward a broader, integrated strategy for managing financial risks.
One key to making this paradigm shift work is Adaptive Behavioral Analytics, which monitors and learns from customer activity, thereby reducing the deluge of false positives that have long frustrated AML teams by offering a better way to identify risk and the associated financial crimes.
To learn more, have a look at the video below, which outlines the AML capabilities of Featurespace’s ARIC™ Risk Hub:
Share