The future of AML? Artificial intelligence will predict suspicious transactions
It has only been a few years since the US Federal Reserve imposed a record fine of 5.6 billion crowns on French bank BNP Paribas for inadequately policing traders who colluded with competitors to buy and sell currencies. The shortcomings in the supervision and internal control of traders served as a warning finger to other banks and companies and a memento of how crucial it is not to underestimate AML. For when a problem does occur and the authorities do decide to apply it, it may have a far-reaching impact...
[.infobox][.infobox-heading]Three letters essential to your business[.infobox-heading]The abbreviation AML stands for Anti Money Laundering, i.e., measures (controls) against money laundering through financial institutions. So-called AML measures must be adopted by all persons who carry out certain activities listed in the AML Act. In addition to credit and financial institutions, this includes persons providing services related to virtual assets, legal entities that manage investors’ assets, second-hand goods traders, gambling operators, and other entities. AML obligations include, for example, client identification and screening, verification of international sanctions, data retention, confidentiality obligations, and many more.[.infobox]
Keep an eye on people and data
At Trask, our AML experts have been helping (not only) banks for several years. What does it look like in practice? “The AML software tools we provide to banks keep track of who their clients are. This is known as KYC (know your client). The second dimension is that individual transactions are checked using software tools,” explains Milan Bartoš, Trask's manager.
- Firstly, the clients of banks and insurance companies are screened to ensure that they are not part of the mafia or organized crime and that they are not politically exposed persons (because of corruption). The “vetting” process is therefore deepest at the time of client onboarding, but is nonetheless continuous.
- The second and more important dimension is transaction monitoring. Automation can effectively help with this, saving time and human labor. Reliable software takes care of everything. “Trask started with AML about seven years ago, when a team of experts implemented the SAS solution for MONETA Money Bank. It is still a well-maintained tool there. A few years ago, we also created our own tool, which is successfully running in Komerční banka’s insurance company and Partners bank,” Milan points out. If AML is to work, he says, it needs real-time data at a reasonable price. Furthermore, the solution must be built on modern streaming technologies that will survive the next few decades.
The future is now
In the future, AI needs to learn its work steps from a sample of activities that are normally done by an investigator examining transactions (and whose know-how is top secret) and to create and predict scenarios defined in the software itself.
“This is the direction that modern AML tools now need to go to make teams' jobs easier. There is more work due to the more efficient detection of suspicious transactions and it is necessary to work faster. The robot will therefore simulate the work of an investigator. And it would be nice to prepare for that. We can help too. Everything will revolve around the existing AML tools - something that already exists will be applied, but we will transfer it to another domain,” Milan predicts.
The work of the investigator will be simulated by a robot.
‒ Milan Bartos
Thus, the tools will run in robotic automated mode and AI will take command “inside”, manipulating (or creating) detection scenarios.
Don’t wait for a fine. If you are interested in learning more about our AML solution, don’t hesitate to get in touch with our Trask's experts. We will be happy to help you automate the controls that are critical to your business.
