A leading argument for why financial institutions have not rushed to adapt or implement artificial intelligence (AI) into their current anti-money laundering (AML) compliance programs is the looming fear over how the regulators would react to the use of this emerging technology.

To take the guess-work out of compliance, the discussion needs to shift from “why” the regulators should get on board with financial institutions using AI, to “how” AI can be implemented in a way that makes the regulators more comfortable and is complementary to current AML processes.

While Know Your Customer (KYC) is an important tenet for effective AML compliance, the concept of Know Your Data (KYD) is emerging as a significant aspect for compliance programs. Financial institutions are plagued with a deluge of data from various sources ranging from retail, correspondent banking, trade finance, securities, KYC, and many others.

Since data is stored in a wide variety of formats, it’s almost impossible to effectively utilize and process data in transaction monitoring systems (TMS). Advanced data science techniques such as AI and machine learning can assist financial institutions in triaging their data for more efficient and effective TMS processing.

AML practitioners have long understood that approximately 95 percent of the alerts that clear a TMS are false positives, which are actually normal, licit transactions of legitimate clients. This alarming statistic requires enormous human investigative effort to properly review all anomalous transactions.

Unfortunately, AML investigators are hamstrung by a limited investigative window of 45 minutes per case (on average) and dangerously outdated technological resources, such as rules-based TMS. Beyond the technological and physical time restraints, human investigators also bring different levels of skill, training, and biases to the job.

However, the process of data cleansing and enrichment with AI-based solutions can significantly help banks and other financial institutions reduce the alarming rates of false positives, while simultaneously keeping the regulators happy.

Instead of applying AI analysis to transactions post-TMS, the forward-thinking approach is to implement AI and machine learning components pre-TMS. New solutions are available today that allow financial institutions to cleanse their data through a comprehensive entity resolution and entity verification process; thereby, cleansing and normalizing the data before it enters a TMS.

By cleansing the various lines of business data sources pre-TMS, AML compliance teams will have the ability to populate the TMS with more accurate and enriched data. AI and machine learning, utilized pre-TMS, will search for entity and transaction commonalties and merge previously duplicated entries prior to analysis.

Artificial intelligence technology has progressed to a point where regulators should be comfortable with using the technology themselves as they benefit from having a unique system-wide view of all covered institutions.

On that note, a new AML modernization bill was recently proposed which promises to bring AI and other new technologies to the forefront. The legislation would “improve FinCEN’s administrative rulings process and require Treasury to explore the potential for AI, machine learning, and other technologies to help detect and prevent money laundering and terrorist financing.”

Recent Posts

What’s Driving AML Compliance Transformation in 2022 and Beyond?

A new report by Aite-Novarica (Aite) examines what’s driving transformation in anti-money laundering (AML) compliance. Specifically, the impact report examines the current AML ecosystem, key trends impacting financial institutions (FIs) and their AML compliance functions, and how they invest in technology and innovation to tackle today’s ever-evolving risk landscape.

read more