Why Do We Need New Customer AML Risk Rating Models?

5 mins

Estimated at between US$800 billion to US$2 trillion every year, money laundering is a serious problem for the global economy. While regulators and financial institutions are working hard to prevent and reduce the crime, launderers are becoming increasingly sophisticated and are introducing techniques that are harder to decode. Changing customer behaviour and the introduction of numerous digital transaction methods add to the AML compliance worry of banks and impact their customer risk rating models.

The COVID-19 pandemic has also been playing its role, as criminals adapt their strategies to the unforeseen situation. We have previously written about the rising number of cybercrimes and fraud schemes across the globe, where criminals take advantage of the people’s fear, helplessness, the need for immediate financial assistance and medical supplies among others.

AML compliance failures are seemingly on the rise as AML fines in the first six months of 2020 reached US$706 million, up from US$444 million in the entire 2019, according to research. It was also revealed that customer due diligence (CDD), AML management, suspicious activity monitoring and compliance monitoring and oversight are the areas where firms are going wrong repeatedly.

What is needed is a new AML risk rating approach, powered by modern technologies such as AI and machine learning. Tookitaki has developed various solutions in relation to customer due diligence, transaction monitoring and screening. Here, we will focus our innovation in the area of customer risk scoring which is one of the primary tools for Know Your Customer (KYC), CDD and enhanced due diligence (EDD) and continuous monitoring of customers.

The Importance of Customer Risk Scoring

Before onboarding customers, financial institutions are mandated to assess AML risk related to them based on a number of factors such as occupation, income sources, and the banking products used. They conduct customer due diligence and monitor the risk ratings throughout a customer’s lifecycle to make informed decisions on potential money laundering cases.

Banks usually do an identity verification and risk assessment for their individual and corporate customers by collecting various details about them. The process is to ensure that they are not doing business with people or institutions involved in financial crimes such as money laundering and terrorist financing. Banks collect as much data as they can about their customers, analyse the data they obtained, determine their risk and provide a risk rating.

Customers with a high risk rating are closely monitored for their actions. Low-risk customers are also monitored but not as diligently as high-risk customers. Even after onboarding a customer, banks periodically update their database about customers. Typically, they do data updates for high-risk customers more frequently than low-risk customers.

Pitfalls of The Current Customer Risk Rating Matrix

Many of the current customer risk rating models are not robust to capture the complexities of modern-day customer risk management. Customer risk ratings are either carried out manually or are based on matrices that use a limited set of pre-defined risk parameters. This leads to inadequate coverage of risk factors which vary in number and weightage from customer to customer.

Furthermore, the information for most of these risk parameters is static and collected when an account is opened. Often, information about customers is not updated in the required format and frequency. The current models do not consider all the touchpoints of a customer’s activity map and inaccurately score customers, failing to detect some high-risk customers and often misclassifying thousands of low-risk customers as high risk.

Misclassification of customer risk leads to unnecessary case reviews, resulting in high costs and customer dissatisfaction. Adding to this, the static nature of the risk parameters fails to capture the changing behaviour of customers and dynamically adjust the risk ratings, exposing financial institutions to emerging threats.


The AI Way of Creating an AML Risk Assessment Matrix

Today, modern technologies like AI and machine learning are getting widespread attention for their ability to improve business processes and regulators are encouraging banks to adopt innovative approaches to combat money laundering. In the area of customer risk scoring, their  is a need for more sophisticated technology that can capture the complete customer activity through proper identification of risk indicators and continuously update customer profiles as underlying activities change.

Keeping that in mind, we have developed a Customer Risk Scoring (CRS) solution as one of the modules of our award-winning Anti-Money Laundering Suite (AMLS). Powered by advanced machine learning, the module addresses the market needs and provides an effective and scalable customer risk rating solution by dynamically identifying relevant risk indicators across a customer’s activity map and scoring customers into three risk tiers – High, Moderate and Low.

The solution adapts to changing customer behaviour to build a 360-degree risk profile thereby providing a risk-based approach to client management. It comes with a powerful analytics layer that includes actionable insights and easy explanations for business users to make faster and more informed decisions.

The key benefits of our CRS solution are:

Broader risk coverage:  CRS assesses risk across a comprehensive range of risk indicators that provides a 360-degree view of AML Risk relative to the customer, their relationships and activities. These dimensions are Customer, Counterparty, Transactions and Network Relationships.

Dynamic customer assessment: The solution provides continuous, on demand and accurate customer risk scoring. Customer AML risk assessment adapts over time to actual customer behaviour. This vastly reduces false signals and improves inappropriate behaviour detection. In short, it acts as a perpetual KYC platform for ongoing due diligence.

Solution level agility: The solution is not a single “model”.  CRS has been developed with advanced ongoing self-learning to evolve based on what is happening within specific client portfolios, business policies and industry trends. This functionality is controlled by client configuration to support all model governance policy and regulation requirements.

Accelerated risk assessment: CRS filters and presents the most critical information needed for investigators to make effective risk-based decisions timely and consistently. The solution simplifies highly complex machine learning decisions into understandable and actionable information.

Reduced time-to-value and clear migration path to ML-based workflows: CRS does not require time and cost consuming change of existing Customer Risk Policies and Controls. Initially complementing your legacy operating environment, CRS provides the functionality required to transition to full machine learning-based AML Customer Risk as and when it is appropriate.

Reduced cost of compliance and reputational risk: The solution helps identify high-risk customers and enable banks to take proactive measures to mitigate the risk of financial loss due to penalties along with various other regulatory, legal and reputational risks.

Money laundering across the globe has increased not just in volume, but also in terms of complexity and sophistication. Customer behaviour has significantly changed with digital banking, transferring funds across geographies has become very easy and even instant in some cases. Such transformations make financial institutions more vigilant. They need to continuously evaluate their customers’ risk score based on their behaviour and monitor based on their updated risks at all times.

As regulators are becoming more stringent globally around AML compliance, strengthening the AML systems continues to remain among top priorities. Our CRS solution enables financial institutions to realise benefits with dynamic customer risk scoring, leveraging advanced machine learning models for improved effectiveness of Enhanced Due Diligence with fewer resources.

To learn more about our AML solution and its unique features, contact us and we will be happy to give you a detailed demo.