trending
In a new white paper, Halyna Hermanns and Dr. Stephan Lemkens from INFORM GmbH outline developments in AI that deliver for both Fraud and AML.
Here, they explain why banks should use a single system in both areas.
AI: a track record of improving Fraud and AML
Over the last decade, Artificial Intelligence (AI) has brought a host of benefits to AML and fraud defence.
Our white paper details how previous waves of AI improved banks’ ability to spot anomalies in transaction data, identify fraudulent behaviour patterns and automate certain fraud detection functions, as well as facilitating compliance with AML rules both nationally and internationally.
One 2020 survey of global banks[1] found 80% of banks agreed that AI has reduced payments fraud, while 63% said AI helped prevent fraud attempts before they happened.
Meanwhile, a new era of generative AI solutions has begun to make its presence felt.
This includes Generative AI, which simplifies and personalizes fraud explanations for each unique case, making it easier for everyone to understand the decision-making process, including the regulator.
Employing Generative AI in fraud detection and AML compliance allows organizations to train their models on a dataset of legal and illegal transactions to learn patterns which can then be used to identify suspicious activities.
“Recent developments in AI make a single shared technology for Fraud and AML defences possible.”
At INFORM, we believe Fraud and AML teams should use a single technology for both functions – and recent developments make this possible, as we set out in our paper.
By using shared resources, financial institutions strengthen their overall risk management thanks to improved collaboration and communication.
They also improve detection rates, speed up case handling and reduce the requirement for manual engagement, creating greater operational efficiency.
Specific AI typologies that improve performance include:
Data-driven AI and Machine Learning (ML) uses algorithms to detect patterns of fraudulent or risky activity, even if those patterns are previously unknown. ML can also create Transaction Risk Analysis scores to exempt transactions from escalated authentications and reduce friction for clients during transactions.
Knowledge-based Expert Systems permit new regulatory requirements and rules such as the forthcoming AMLA global rule book or the EU’s DORA to be rapidly implemented thanks to efficient AI that enables the processing of large amounts of data in real time.
Hybrid AI facilitates more profound, real-time analyses of transactional data, mitigating risk during high-speed situations such as instant payments. Instead of relying solely on data-driven machine learning methods, Hybrid AI technology blends machine learning with knowledge-driven techniques such as fuzzy logic-based scorecards and watch lists as well as dynamic profiling. This improves fraud detection even where data is complex or imprecise.
Our white paper sets out how sharing resources between fraud and AML gives banks a 360-degree view of all their customer engagements, rather than just seeing transaction information related to possible fraud or compliance rule breaches.
Costs are reduced in a number of areas, from reduced full-time employee (FTE) costs thanks to a shared workload, and lower training costs associated with fewer FTEs.
Dealing with just one system also means that operational costs are reduced.
Furthermore, the use of one AI-driven system lets banks strengthen risk management across the board by becoming more responsive to dynamic risks.
The implementation of new regulatory requirements and rules is faster and more accurate, as are fraud detection and the identification of AML violations.
Finally, sharing information through one resource helps teams identify patterns in AML breaches and fraudulent behaviour and makes problem-solving more effective, including the rapid identification of new fraud types, changes in customer behaviours and more.
At a time when banks are caught between investor pressure to improve revenue and margins on the one hand and the need to deliver improved fraud and AML performance on the other, creating shared resources between Fraud and AML teams makes sound business sense.
The paper demonstrates the benefits of these shared solutions, from reduced costs to improved performance, and also how they open the door to enhanced innovation between internal Fraud and AML teams.
For more on how shared resources benefit your Fraud and AML functions, download the white paper HERE
[1] See PYMNTS, 1 December 2020, “How AI and ML improve fraud detection”
Comments