Institute Insights - Ready for take-off - GenerativeAI Copilots for Financial Crime Prevention

Ready for take-off - GenerativeAI Copilots for Financial Crime Prevention

“Life’s better with company; everyone needs a co-pilot” — George Clooney

Financial crime prevention software providers are harnessing the power of ChatGPT and Artificial Intelligence (AI) technologies to streamline compliance risk activities. At Lucinity, we proudly introduced our very own Luci copilot in May 2023. These innovative solutions, whether by increasing automation or coaching analysts to conduct more effective investigations and analysis, have undoubtedly found multiple applications within the compliance space.

Generative AI tools offer numerous advantages that can greatly benefit various areas:

• Time Savings: These tools significantly reduce processing time for tasks like data query, retrieval, integration, and overall investigation.
• Enhanced Customer Intelligence: By contextualizing customer profiles and presenting them as a cohesive story, generative AI enables a better understanding of customer behavior.
• Consistency: AML analysts are guided to follow a structured sequence when constructing cases and Suspicious Activity Report (SAR) narratives, ensuring consistency in their work.
• Coaching: These tools provide valuable assistance in standardizing processes, defining terms, interpreting data, and identifying typologies, empowering analysts with expert guidance.
• Minimizes Disruption: With generative AI tools such as Luci, work can be performed seamlessly without disruption. For example, Luci provides on-demand updates, eliminating the need for users to halt ongoing investigations.

It's evident that Artificial Intelligence (AI) is steadily progressing beyond its initial role in Transaction Monitoring (TM). This reaffirms three important ideas: firstly, AI is here to stay; secondly, regulations must evolve, catch up, or be created to keep pace; and thirdly, remarkable tools like Luci have the potential to revolutionize not just efficiency, but also foster innovation in integration, connectivity, and overall user experience in AML compliance software.

Achieving Optimal Efficiency: A Challenging Goal

Throughout the history of computer technology and automated solutions in business, there has been a continuous pursuit of enhancing efficiency and discovering those increasingly rare areas where machines can provide an advantage over traditionally human-handled processes.

In the realm of compliance, crucial functions like KYC (Know Your Customer) and Transaction Monitoring have witnessed a remarkable revolution in terms of processing speed. This progress has been matched by the ongoing development of financial crime patterns that consistently manage to stay a step ahead of existing systems and processes.

The quest for optimal efficiency has primarily focused on the quantitative aspects of the monitoring process. For instance, efforts have been directed towards processing a higher volume of transaction monitoring alerts and effectively differentiating true alerts from the noise generated by false positives. Additionally, the use of Robotic Process Automation (RPA) has been employed to streamline complex template filling, thereby facilitating downstream analytical work.

Generative AI changes the rules of the game

GenAI technology has unlocked new possibilities by replicating qualitative capabilities that were previously unattainable. This means that algorithms can now be trained to write content or narratives about cases or Suspicious Activity Reports (SARs).

Taking the example of the UK SAR regime, statistics from 2022 reveal nearly one million reports submitted by financial institutions and other regulated entities. As detection tools generate increasing amounts of data for this process, it's understandable how reporting entities can feel overwhelmed in meeting their regulatory obligations. At the same time, financial intelligence units and law enforcement agencies receiving these reports may lack the necessary resources to effectively process and prioritize such volumes.

With GenAI, reporting units can significantly enhance their investigation processes, operating in a more agile and efficient manner. By utilizing GenAI tools, instant updates on ongoing cases can be provided to brief senior management, minimizing disruptions. These tools can also generate case narratives and draft structured reports, ensuring essential elements are included or flagged if missing.

Moreover, GenAI can improve the quality of SARs as an output, enabling better structure and understanding. By integrating with relevant systems, GenAI can offer instant checks for screening, negative news (adverse media), KYC reviews, Open Source Intelligence (OSINT), risk assessments, and even facilitate seamless communication among team members without the need for data transfers or screen changes. These capabilities not only expedite the process but also provide much-needed context in an environment where criminals exploit ambiguities.

Looking ahead, we envision a future where a data-sharing framework powered by GenAI could bring transparency to areas where criminals thrive. Imagine how a framework like Singapore's Collaborative Sharing of ML/TF Information & Cases (COSMIC) could evolve with the assistance of GenAI, shedding light on illicit activities and promoting collaboration.

With great power comes great [regulatory] responsibility

Before we entertain any alarming notions of a machine takeover, it is imperative to emphasize that the aforementioned advancements should be implemented with a comprehensive regulatory framework in place. This framework necessitates a closer collaboration between regulatory entities, financial institutions (FIs), and software solution providers.

Similar to other areas within Artificial Intelligence, GenAI must adhere to and be integrated into an ethics framework. This ensures that it fulfills its intended purposes and operates within clearly defined boundaries.

Furthermore, it is crucial to closely monitor the operational use of GenAI to prevent what is known as the "online training effect." This effect occurs when users mindlessly click through successive slides or screens solely to reach the end of a lesson, compromising the quality and analysis. To address this concern, a robust system of analytics and auditing tools must support the regulatory architecture.

Conclusion

• Compliance Copilots are undoubtedly here to stay. Their capabilities hold immense potential, and when utilized effectively, they can bring significant benefits that have the power to drive reform in crucial areas such as the investigations process and SAR regime.
• Regulatory bodies must proactively keep pace with the advancements in GenAI. Its implications go beyond what we typically associate with "traditional" technology, encompassing legal and ethical considerations that necessitate attention.
• The application of GenAI is versatile, spanning across various functions, from Product, Engineering & Innovation roles to commercial areas. Its analytical capabilities can be leveraged to enhance product offerings and provide valuable insights to customers.