NArtificial intelligence (AI) and machine learning have been talked about a lot lately. “Artificial intelligence is the field of developing computers and robots that can behave in ways that mimic or exceed human capabilities. AI-enabled programs can analyze data to provide contextual information or automatically trigger actions without human intervention.”1 Simply put, AI is the process by which computers and systems perform tasks in a similar way to humans, but with the speed and decision-making skills defined by humans. Examples of AI in today’s world include smart devices, voice assistants (such as Siri), chatbots, translation applications, navigation/mapping apps, facial recognition tools, and more.
Machine learning is the path to AI. This is a subset of AI that utilizes “algorithms that automatically learn insights from data, recognize patterns, and apply that learning to make better decisions.”2 Basically, machine learning involves some rules and conditions within the AI that learns from huge historical datasets to speed up decision-making.
Organizations are incorporating AI and machine learning into their systems and strategic planning to leverage automation of manual processes and act on data-driven insights more efficiently and effectively.
Financial institutions (FIs) are leveraging AI and machine learning technologies in financial crime compliance, particularly for anti-money laundering/counter-terrorist financing (AML/CTF), sanctions, and customer onboarding, leveraging the benefits described below.
AML/CTF
When it comes to AML/CTF, traditional methods for transaction monitoring systems involve implementing detection scenarios based on specific parameters, logic, and thresholds. This approach is becoming obsolete, and financial institutions are focusing on leveraging service providers who integrate AI and machine learning technologies into their transaction monitoring systems and whose systems are not only based on detection scenario logic and threshold reviews, but also employ smart technologies. In doing so, we go beyond traditional methods and delve deeper into data analysis through various investigative trading patterns of alerted trades and past trades. Analyze customer profiles within a bank’s core banking system. Search for information on social media platforms. Additionally, AI/machine learning technology is customizable, giving organizations the flexibility to select and add additional checks and automatically close relevant alerts as Level 1 at a high pace with a comprehensively customized narrative. The system can be programmed to flag actual cases as Level 2, allowing reviewers to make the necessary decisions accordingly. These technologies are continually learning how to identify complex and anomalous transactions. Some service providers offer the flexibility to automatically close alerts at both Level 1 and Level 2, depending on your organization’s requirements.
sanctions
Many financial institutions face the challenge of a large number of false positives in sanctions review alerts related to customer and payment reviews. Payments are highly sensitive and time-sensitive transactions that must be processed within a defined turnaround time. We have seen banks being penalized for processing payments in violation of sanctions-related concerns. Fundamentally, payments are highly confidential as they must be properly vetted to avoid sanctions violations. At the same time, the real-time nature requires rapid processing. Financial institutions have had great difficulty putting the right parameters and screening logic in place.
Despite having built-in upper thresholds, payment screening can generate a large number of false alerts, which not only takes time but also delays the payment processing cycle. To release these payment review alerts effectively and in a timely manner, some service providers are offering advanced AI and machine learning-related technologies that quickly release bulk alerts with well-customized narratives as Level 1 and refer potential matches to reviewers at Level 2 to take necessary actions. Alerts are reviewed by these models, which analyze the parameters and conditions that triggered the alert for all parties involved in the transaction, and reference customer information, historical trends, detailed Google and media search results, and compliance with the organization’s specific policies/requirements. The alert is then closed or escalated to the next level.
These AI models learn from historical data and evolving trends by improving their detection over time. Again, you can customize the configuration according to your FI requirements. This benefits FIs by increasing not only alert levels but also match resolution rates, allowing compliance teams to focus more on potential matches, reducing the overall turnaround time to dismiss related alerts, and freeing up time to focus on other necessary sanctions activities.
Customer onboarding
Another benefit for financial institutions of adopting AI/machine learning capabilities relates to the customer onboarding process. Some organizations use robotic process automation to automate many workflow steps. However, AI can provide faster results for very large datasets of customer due diligence checks for sanctions screening/negative media searches, Google searches, document checks, and other related items. AI allows you to do this quickly with a comprehensive narrative, giving Level 2 reviewers enough time to make informed decisions without delay and manage their time effectively.
The use of AI/machine learning technologies within AML/CTF, sanctions, and customer onboarding provides a comprehensive management information system for good practice analysis based on various identifiers.
Key benefits of leveraging AI/machine learning in financial crime compliance include faster processing, significantly improved data analysis, and improved handling of complex tasks, all of which result in effective decision-making and help organizations better manage their resources.
AI/Machine Learning can teach these models and tools to perform specific tasks and make decisions in predefined steps involving large datasets and algorithms. The clearer the definitions of these models, the more effective the results will be.
The future of financial crime compliance is AI, and the technology is continually evolving with advanced capabilities. These tools must be used ethically and responsibly.
Ghaus Bin Ikram, CAMS, Head of Financial Crime Compliance/MLRO, United Arab Emirates PJSC, United Arab Emirates 
Disclaimer: The views expressed are solely those of the author and do not represent the views of his employer.
