Banks and insurance companies have to comply with ever more laws and regulations while at the same time criminal activity and data volumes are rocketing. Financial institutions don’t want and can’t afford increased workloads and higher costs. As a result, many are turning to artificial intelligence to help them harness automation to reduce the burden on staff and cap costs.
Why Compliance Is Now So Critical in the Financial Sector, and How Machine Learning Can Help
Sharp rise in money laundering reports is a challenge for financial regulators and companies
The increase in criminal activity – partly fuelled by the Covid pandemic – is piling the pressure on compliance departments and regulators alike. For example, the volume of suspicious activity reports related to Germany’s Money Laundering Act increased more than twelvefold between 2010 and 2020. The sharp 25% increase in 2020 is also a result of news related to COVID-19 and cryptocurrencies. Other sectors such as notaries also now have to comply in this respect and are required to report transactions linked to money laundering. The FIU in Germany forecasts that the number of SARs will be higher still in 2021.
The situation is similar in other countries: in 2020, the Money Laundering Reporting Office Switzerland (MROS) received more SARs than ever before. The SFIU in Liechtenstein also says volumes are soaring, while the FIU in the Netherlands recorded a 163% increase in the number of suspicious transactions in 2020. It expects the numbers to keep growing, which is why it joined forces with IT specialists and data scientists to set up a DevOps team in 2020.
Flood of legislation ramps up the pressure on finance and insurance industry
The financial sector has been flooded with new legislation over the last few years. Key regulations for compliance departments include the EU Money Laundering Directive, the Market Abuse Directive (MAD) and MaRisk compliance. 2020 saw an additional raft of legislation being introduced in Germany, including anti-money laundering regulations based on the all-crimes approach, BaFin’s AUA guidelines relating to the Money Laundering Act, and the European Transparency Register and Financial Information Act (TraFinG). Switzerland has also amended certain legislation, and the revised Money Laundering Act will come into force in 2022. In 2024, the EU’s new Anti-Money Laundering Authority (AMLA) is set to begin work on creating a single integrated system to combat money laundering and terrorist financing. All of this means more work for banks and insurers.
Compliance is now a critical issue in the financial sector
This has all increased the pressure on compliance teams in their fight against money laundering – exacerbated still further by the shortage of skilled staff and rising costs. Banks and insurance companies have already invested massively in compliance, but conventional methods are simply inadequate to tackle the increasingly complex work involved and the growing flood of data. Yet if internal control mechanisms fail to do their job, banks risk incurring massive fines that can run to millions. Banks all around the globe were hard hit by this.
False positives as a cost driver
Staff shortages in compliance departments and rising costs are ramping up the pressure on banks and insurers. One step they can take is to reduce false positives, also known as false alarms. Rule-based IT systems generate these reports, which are technically accurate but on closer inspection clearly do not represent a money laundering risk. They place an unnecessary burden on compliance staff because every false positive has to be clarified. Many financial institutions have already tackled this issue and achieved real improvements thanks to machine learning.
Why is machine learning so prominent?
Machine learning – a component of artificial intelligence – analyzes potential money laundering cases based on data knowledge. Supplemented by the knowledge and experience of compliance staff in the specific area, it is possible to identify anomalies with greater speed and efficiency. As a result, huge data volumes can be analyzed more efficiently, suspicious patterns detected more easily, and potential risks identified at an early stage. This is also the conclusion of a recent study by the FATF (Financial Action Task Force on Money Laundering), a leading international body in the fight against money laundering. By using automation and machine learning, compliance departments can reduce the amount of work involved in complex analysis and review, increase efficiency, and cut costs.
In practice, banks are shrinking their workload by 50%, for example in payment screening. Another example is PEP and sanctions list screening. Once again, machine learning is helping to reduce hit rates by up to 60%, allowing compliance teams to focus on true positives and allocate their time accordingly.
Whitepaper: “Why successful banks now rely on machine learning in compliance.”
The whitepaper draws on recent research to highlight the latest trends and success factors for compliance departments in the financial sector. It also explains more about how digitalization and machine learning works and outlines the benefits of using machine learning.
Would you like to know more? Download the whitepaper here:
These might be of interest to you
Machine learning in use: Optimized fuzzy payment screening at VP Bank Group
A new payment screening software aided by artificial intelligence facilitates VP Bank Group to improve the quality of its hits with 50% less workload. The success story shows how efficiently machine learning can reduce the number of false positives.Learn more
Video: How AI helps to improve embargo checks in payment screening
In this short video our colleague explains and illustrates important steps for payment screening with artificial intelligence. Enjoy the vivid explanations and dig deeper into the topic.Watch the video
Better compliance thanks to digitalization: Why successful banks now rely on machine learning
The white paper provides an overview of the latest SAR figures, new regulations, and key steps towards greater automation in Compliance. It also includes 9 lessons learned when integrating machine learning into compliance systems.