The new horizon for anti-money laundering solutions

Evolve to scalable AI-powered set ups

Publish date:

Data and AI solutions to boost anti-money laundering robustness and efficiency are out there. So why are financial institutions still spending billions to combat financial crime?

Financial institutions are being squeezed from both sides in the crusade against money laundering. On one side are regulators and the risk of serious fines and investigations. On the other are valued customers who will only tolerate so many invasive and time-consuming questions before they move their accounts. And in the middle: overworked anti-money laundering (AML) staff, tasked with sorting, investigating, and reporting suspicious potential money laundering activities on a scale never seen before.

But what’s striking about this situation is that solutions already exist – multiple solutions that can dramatically increase the accuracy of AML detection systems. Financial institutions are aware of them – sometimes they even develop them. So, what’s holding them back?

The state of AML today

In the arms race between money launderers and the financial institutions trying to stop them, agility is key. As criminal organizations find new ways to slip their funds through the AML nets, institutions must continuously notice the activity, flag it, investigate it, and then update their models to accommodate the new information. Institutions are keeping pace, but the costs generated by this process take an astounding toll. The total global cost borne by financial institutions to combat financial crime fell just short of $214 billion in 2020. US banks spent between 0.4% and 2.4% of their total 2018 operating expenses on AML and Bank Secrecy Act compliance, breaking down to between about $5 and $44 for each account. Errors can be financially devastating. Since 2009, financial firms around the world have paid out a total of $40 billion in regulatory fines for non-AML/KYC compliance, including over $10 billion in 2020 alone. The problem is exacerbated by the growing complexity of finance in general and the sheer volume of transactions. The data has become too much to handle efficiently, as seen in the false positive rate of 95%. When institutions lose sight of their data, they lose their agility. This risks giving criminals the upper hand.

The need for scalable AML data platforms

The inability to rapidly access large amounts of data lies at the root of the problem. Most financial institutions today rely on a data architecture that wasn’t built to handle the volume of data and analytics that AML demands. There are multiple market solutions available now that support smart analytics applied to AML – these include major editor solutions, data analytics solutions (some of which come with embedded AML-specific features); even giant cloud service providers are beginning to get involved. These market solutions bring substantial savings by linking disparate data points, drawing conclusions, and automatically updating ML models. They are also beginning to include more and more features that probe potential risks, leveraging advanced AI/ML models. But these technologies depend on unfettered access to data across institutions. They cannot work with scattered data storage and complex data exchange rules that often require internal approvals. Next-gen AML solutions simply can’t run on the last generation’s data architecture.

Siloed solutions

One case that exemplifies the problems caused by aging data architecture is the inability to scale successful solutions. There’s been an explosion of creative solutions in recent years, as AML teams pool their information and devise improved detection systems. The problem is that these solutions usually don’t make it out of the POC/POV mode. They’re rarely scalable, or even replicable. When the small team of data scientists who designed the solution split up or move on, the solution often goes too. An immense amount of value has been lost in this way. Employees are continuing to explore value-adding use cases, solving problems as they come and sometimes creating some structurally brilliant solutions. But, in all likelihood, these solutions will be lost too. Without a robust data architecture, they’re falling on barren soil.

Strategies for here and now

As new data architecture comes to financial institutions, it will inevitably help combat money laundering. But firms don’t need to wait for an entirely new data management system to start benefiting from next-generation AML. The first step is to find ways to effectively leverage existing capabilities, evolving from tactical POC/POV to industrial data setups. It is possible to increase robustness and efficiency in risk detection now, bringing immediate benefits and creating the flexibility to drive continuous adaptation in a world of ever-changing money laundering threats. In fact, much of the current data transformation occurring in financial institutions is driven by the requirements of AI, which overlap substantially with AML. This creates opportunity. With the right approach, financial leaders can change the face of AML today.

To learn more about the steps you can take to enhance your AML capabilities now, contact mathias.ros@capgemini.com or jean-charles.croiger@capgemini.com.

Authors

Jean-Charles Croiger Mathias Ros

Jean-Charles Croiger

Director – Financial Services Compliance powered by Data, Capgemini Invent

Mathias Ros

Engagement Manager, KYC/AML Solutions and Platforms, Insights and Data, Capgemini