Skip to main content
Financial Crime Compliance Predictions 2020: More AI and Robots

Behind the headlines on financial crime compliance are big challenges. Criminals are getting more and more sophisticated. Rapid payments, instant payments, PSD2 have been vehicles to further misuse the financial system for laundering money and harming banks, their customers and the economy worldwide.

On the other side, there is a lot pressure on banks to stop the constantly increasing costs of being compliant. Many institutions expanded the size of their compliance headcount by over 500% in the past few years. North American financial services firms nowadays spend more than $31.5 billion a year on ensuring AML compliance.

Financial crime compliance has become a real cost burden for most banks in times of shrinking profits. Hence, the focus has shifted for many institutions on operationalizing compliance to become more efficient and effective at the same time.

Here’s what we see for next year:

Prediction 1 – More AI

We see a trend and a mind shift of regulators which will help financial institutions. Regulators are more open to new methods like the use of AI (artificial intelligence), machine learning and robotics. In fact, they are actively encouraging banks to consider, evaluate and, where appropriate, implement these innovative technologies.

I have asked many regulators if they would allow financial institutions to fulfill regulatory requirements with the use of AI. The answer has been: That would match the expectations, as long as the AI provides a proper explanation as to WHY an alert was generated.” That leads us into the topic of explainable AI. Reason codes that business managers can understand help investigators and satisfy regulators.

This trend does not mean we throw away the existing risk-based approach, which is based on a good compliance knowledge in defining “detection scenarios” (e.g., based on FATF2012) with rules. What we see is a co-existence – a hybrid of the existing scenarios and the AI mechanisms. This will help:

  1. Prioritizing the scenario-based alerts
  2. Quickly and automatically adopting on new money laundering schemes.

Prediction 2 – More Robots

Robotic process automation allows financial institutions to streamline and automate the process of investigation and alert handling in KYC and AML. Nothing can be more boring and expensive than hiring armies of investigators to simply close false positive alerts, which typically range between 75-90%.

Repeatable manual tasks are typically

  • High-volume
  • Time-consuming
  • Prone to error
  • Based on clearly defined rules and criteria

That is what a machine — a robot — can do much better. Clearly defined alert rules and case rules (which are specific to the situation of the financial institution, the products, the customers, etc.) will take away certain repeatable manual tasks from a user. Investigators can thus focus their activity on the remaining more complex tasks. Robotics should be integrated into an enterprise-wide alert and case management. We foresee that the next version of anti-financial crime solutions will provide a large variety of capabilities in this area.

Analytics-driven alert prioritization and RPA helps lower costs in the short term. But the combination of both can also lead to enormous effectiveness and efficiency gains. Our experiences have shown that these technologies can increase the number of SARs by 20% while at the same time producing efficiency gains of 30% in alert and case management.

Prediction 3: Fraud and Financial Crime Management Convergence

Another way financial institutions will look to lower costs and improve results next year is through the convergence of AML and fraud detection. The systems used today for these functions are similar and fulfil many common requirements, such as detecting unusual behavior — but banks still operate them in a siloed manner. In their recent survey Ovum found that more than 80% of financial institutions want to achieve the goal of breaking down the siloes from an organizational point of view within the next 8 years.

Using a fully scalable IT environment fulfilling both the requirements for detecting fraud and AML at a time not only provides an economy of scale but would also allow institutions to take a “cross border” view to detecting illicit activities.

related posts