The product aims to provide a suitable alternative to rules-based transaction alerting by offering a consolidated machine learning (ML)-generated customer risk score. This score ties in with the bank’s data, which includes network behaviour, transactional patterns, and Know Your Customer (KYC) data to identify instances and groups of high-risk retail and commercial customers.
The product is also capable of adapting to changes in underlying data to offer more accurate results that ultimately enhance operational efficiency. Some of the most noteworthy technologies behind this product are proprietary machine learning technologies as well as Google Cloud systems such as Vertex AI and BigQuery. In order to help financial institutions to expedite the investigation and improve customer experience, Google’s offering handles the complexities of running ML at scale and offers detailed explanations of the outputs.
According to the official press release, this new Google offering is capable of enhancing customer experience by improving precision and reducing false positives. In essence, AML AI reduces the need to engage with customers for additional compliance verification checks.
The system can also minimise wasted investigation time by reducing alert volumes and offering explainable outputs. It can detect instances of financial crime risk with accuracy, and it provides financial institutions with auditable and explainable outputs to support internal risk management.
The global impact of money laundering
The official press release details that the amount of money laundered each year is estimated to reach 2 to 5% of the global GDP, or up to USD 2 trillion annually. In many cases, the proceeds gained from money laundering are linked to illegal activities such as drug and human trafficking, as well as the financing of terrorism.
Google’s AML AI launch aims to address the time-consuming and resource-intensive natures of current anti-money laundering programmes, which rely on manually defined rules. The problem with rules-based systems is that money launderers can learn to work around the rules in order to avoid detection. The press release further reveals that more than 95% of system-generated alerts turn out to be false positives in the first phase of the review, with approximately 98% never ending up in a suspicious activity report (SAR).
These high false positive rates are addressed with manual reviews that can sometimes distract institutions from true suspicious activity.
Every day we send out a free e-mail with the most important headlines of the last 24 hours.
Subscribe now