Talk to an Expert
Talk to an Expert
Talk to an Expert
Talk to an Expert

        Case Study

        Transforming Legacy Compliance System with Next-gen Technology for a Universal Bank

        Bank-UOB

        THE CLIENT

        A leading Singapore-based Universal Bank

        United Overseas Bank (UOB) is a leading bank in Asia with a global network of more than 500 offices and territories in Asia Pacific, Europe and North America. 

        UOB has a strong risk-focused culture using next-generation technologies to stay vigilant in an ever-changing financial crime landscape. With the commitment to enhance its AML surveillance, UOB saw a significant opportunity to tap into machine learning to augment and enhance its existing systems to spot and prevent illicit money flows.

        THE CHALLENGE

        Implementing a True Risk-based Approach and Reducing the Rising Cost of Compliance

        Risk

        UOB faced a strategic imperative to optimise its alert management and address the rising cost of compliance. With the volume and velocity of transactions flowing through the bank, there was a pressing need to reduce "false positives" and efficiently close alerts.  

        UOB was also keen to acquire better insights from the transactions and activities of high-risk individuals and companies and suspicious activities to remain vigilant against any potential money laundering activities. Having experimented with multiple systems, UOB faced challenges in finding a sustainable solution.

        1
        1
        High Transaction Volume

        Coping with the sheer volume and velocity of transactions.

        2
        2
        False Alert Challenges

        Overcoming a staggering 90-95% false alert rate across transaction monitoring and customer name screening.

        3
        3
        Missed High-Priority Cases

        Ensuring that no high-priority cases were overlooked in the deluge of alerts.

        THE SOLUTION

        A Future-ready ‘Self-learning’ Machine Learning Solution  

        Fincense Illustration-1

        UOB embarked on a transformative journey by collaborating with Tookitaki, to integrate machine learning (ML) into its anti-money laundering program. This partnership aimed to propel UOB into a future-ready 'Community-driven compliance model.' At the core of this initiative was the deployment of Tookitaki's Anti-Money Laundering Suite (AMLS) to transform the transaction monitoring and name-screening process.

        Tookitaki AMLS Smart Alert Management

        Tookitaki deployed its proven Smart Alert Management solutions to transform the current system for transaction monitoring and name screening. AMLS Smart Alert Management (SAM) leverages a combined supervised and unsupervised ML techniques, to swiftly detect suspicious activities and pinpoint high-risk clients with enhanced accuracy. The key components of this solution included:

        • Seamless Integration: AMLS employs standardized data schema and adapters to integrate with the legacy systems
        • Risk Classification: AMLS excelled in classifying AML risk, demonstrating precision through L1-L3 buckets, and maintaining an accuracy rate exceeding 85%.
        • Adapting to Skewed Data Sets: During COVID-19, the alert data showed skewness due to high defensive reporting. However, AMLS showcased resilience, by adapting to the skewness and delivering consistent results.
        • Reduction in False Positives: SAM showed increased effectiveness in identifying suspicious patterns and reduced false positives by 50%-70%
        Screening SAM

        Innovative Features of AMLS Smart Alert Management

        Tookitaki AMLS harnesses the robust capabilities of our advanced machine learning platform, Tookitaki Data Science Studio (TDSS). TDSS is equipped with prebuilt machine learning pipelines that autonomously construct models tailored to meet diverse requirements across all AMLS modules. 

        Automated Model Management

        The Automated Model Management framework automatically constructs a Machine Learning model trained on client data, delivering a bespoke model, instead a generic industry ML solution.

        Alert Prioritization Engine

        The Smart Alert Management (SAM) Risk scoring engine employs a dual approach of supervised learning and unsupervised learning to ensure precise alert prioritization. 

        Champion Challenger Framework

        Tookitaki’s machine learning platform features a Champion–Challenger Framework which is a self-learning system designed to automatically adapt to changing customer data and deliver consistent performance.

        Explainable AI Framework

        Cutting-edge Explainable AI (XAI) Framework, currently under patent review, pioneers a 'glass-box' solution for enhanced transparency. It provides the output of the ML model in simple human-readable language for quicker investigation and audit.

        THE RESULTS

        Focused on optimizing the detection of new and unknown suspicious patterns and prioritizing known alerts, UOB witnessed a significant advancement in its transaction monitoring and name screening modules.

        Transaction Monitoring

        0 %
        Increase in True Positives
        0 %
        Reduction in False Positives with less than 1% Misclassification

        Name Screening

        0 %
        Reduction in False Positives for Individual Names
        0 %
        Reduction in False Positives for Corporate Names
        Frame 196

        “The area of AML requires constant  vigilance and continual enhancement. The use of RegTech such as Tookitaki’s AMLS enables us to augment our ability to identify actionable alerts and minimise false positives. These sharpen the accuracy and effectiveness of our AML risk management.”

        Head of Group Compliance at UOB
        jake-overlap-card-img

        Time to reform your compliances

        Kickstart your journey by exploring our products or request a demonstration with us.
        Talk to an Expert
        FinCense Modules
        Customers
        Resources
        General Info