Faced with mounting compliance costs and regulatory pressures, financial institutions are rapidly adopting Artificial Intelligence (AI) solutions, including machine learning and robotic process automation (RPA) to combat sophisticated and evolving financial crimes.
Over one third of financial institutions have deployed machine learning solutions, recognizing that “AI has the potential to improve the financial services industry by aiding with fraud identification, AML transaction monitoring, sanctions screening and know your customer (KYC) checks” (Financier Worldwide Magazine).
When deployed in financial crime management solutions, analytical agents that leverage machine learning can help to reduce false positives, without compromising regulatory or compliance needs.
Challenges of conventional approaches
It is well known that conventional, rules-based fraud detection and AML programs generate large volumes of false positive alerts. In 2018, Forbes reported “With false positive rates sometimes exceeding 90%, something is awry with most banks’ legacy compliance processes to fight financial crimes such as money laundering.”
Such high false positive rates force investigators to waste valuable time and resources working through large alert queues, performing needless investigations, and reconciling disparate data sources to piece together evidence.
The “highly regulated environment makes AML a complex, persistent and expensive challenge for FIs” but increasingly, “AI can help FIs control not only the complexity of their AML provisions, but also the cost” (Financier Worldwide Magazine).
Benefits of machine learning
In an effort to reduce the costs of fraud prevention and BSA/AML compliance efforts, financial institutions should consider AI solutions, including machine learning analytical agents, for their financial crime management programs.
Machine learning agents use mathematical and statistical models to learn from data without being explicitly programmed. Financial institutions can deploy dynamic machine learning solutions to:
- Catch emerging criminal trends by identifying complex fraud and money laundering patterns
- Save time for investigators by filtering routine transactions that cause rules-based systems to generate false positive alerts
More data, more effective training
To effectively identify patterns, machine learning agents must process and train with a large amount of quality data. Institutions should augment data from core banking systems with:
- Ancillary banking systems
- Third-party data
- Open-source data
- Consortium data
When fighting financial crime, a single financial institution may not have enough data to effectively train high-performance analytical agents. By gathering large volumes of properly labeled data in a cloud-based environment, machine learning agents can continuously improve and evolve to accurately detect fraud and money laundering activities, and significantly improve compliance efforts for institutions.
Real results: 66% false positive reduction
Importing and analyzing over a billion transactions every week in our Cloud environment, Verafin’s big data intelligence approach allows us to build, train, and refine a proven library of machine learning agents. Leveraging this immense data set, Verafin’s analytical agents outperform conventional detection analytics, reducing false positives and allowing investigators to focus their efforts on truly suspicious activity. For example:
With proven behavior-based fraud detection capabilities, Verafin’s Deposit Fraud analytics consistently deliver 1-in-7 true positive alerts.
By deploying machine learning, Verafin was able to further improve upon these high-performing analytics – resulting in an additional 66% reduction in false positives. Training our machine learning agents on check returns mapped as true fraud in the Cloud, the Deposit Fraud detection rate improved to 1-in-3 true positive alerts, while maintaining true fraud detection.
These results clearly outline the benefits of applying machine learning analytics to a large data set in a Cloud environment. In today’s complex and costly financial crime landscape, financial institutions should deploy financial crime management solutions with machine learning to significantly reduce false positives, while maintaining regulatory compliance.
In an upcoming article, we will explore how and when robotic process automation can benefit financial crime management solutions.