Bad actors are continuously innovating their money laundering schemes, pushing the compliance organizations within banks and other financial institutions to do the same.
Yet, compliance organizations have struggled in recent years to stem the flow of employee departures and to find candidates who can help overcome shortfalls in employee headcounts. As a result, Level One (L1) compliance team analysts face heavy workloads of repetitive and routine work, leading to high rates of employee burnout. Recent research from Celent found that 70% of banks and non-banking financial institutions face capacity challenges in their compliance operations. This is where artificial intelligence (AI) has the potential to transform the industry.
AI and automation tools are easily accessible and highly effective performing L1 compliance analyst tasks and processes. Moreover, regulators are encouraging financial institutions to innovate and look for new ways to perform old processes via new technologies. This is evident throughout the AML Act of 2020 and is supported by FinCEN’s innovation initiative and the release of the Wolfsberg Group’s principles on AI and machine learning (ML).
AI can be incredibly useful in augmenting your AML compliance program, but it requires planning and well-thought-out guardrails to ensure its efficacy and to avoid running afoul of your organization’s model risk management (MRM) framework. Many AI solutions are black-box models, not open platforms that allow you to fully understand exactly how they work and how their AI makes its decisions. The right AI solutions for banks and other FIs have explainability within the model to ensure that AI is not missing anything and that its decisions are always the right decisions.
How to ensure you leverage AI correctly
The critical first step to perform before selecting and implementing any AI solution is to develop a business case that addresses the potential pitfalls by asking tough questions of your preparedness. While not an exhaustive list, the following questions will help you understand which AI projects to take on and the right solutions for them.
- Do I have the right use case? It is essential that AI, like any other transformative technology, proves its value early. Find a single use case that will provide a fast return in value and that you can deploy quickly.
- What problem am I solving? When you meet with vendors, verify that they are actually delivering against your defined problem at peer banks. It’s a giant red flag if the vendor you have in mind cannot produce solid evidence of prior successes.
- How do I address my stated problem(s) with AI? Familiarize yourself with the best ways to approach AI and ML within your program. Organizations like ACFCS and leading vendors in AI for financial crime and AML compliance have a wealth of knowledge and available advice content to help guide you.
- How am I proving AI will work? Understand and be able to explain to senior management and regulators how your organization will build, test and validate your AI models. Be thoughtful and clear about any new AI solution’s impact.
- Have I arranged for early buy-in from regulators? You will want to understand as early as possible – even during vendor evaluations – all the concerns from each relevant regulator. Recognize the reality that different regulators have varying degrees of knowledge around AI, ranging from practically none at all, to highly conversant in, and supportive of, new AI initiatives.
- Have I arranged to work with impacted teams across the organization? Engage all teams which may be impacted by your new AI solution. These can include modeling teams, data teams, information security teams, IT, and more. Understand what the security strategy is and how AI will work within that.
- How will I align with the MRM framework? Does the MRM framework account for generative AI as well as standard AI decision making? To address this, test, document and thoroughly review your MRM framework to make sure the AI model has gone through the scrutiny needed to be deployed correctly.
- What are my ROI expectations? Harking back to the single use case which you chose, only proceed with it if it can provide a fast return in value and can be deployed quickly. Return can be represented by operational savings and/or revenue support, such as more and faster customer onboarding.
Start your project by answering these questions (or preparing to answer them). If you cannot answer all of them, then you are not ready to proceed. It will be far better to delay for a short period of time and get everything right than to rush in, cause unforeseen problems, and then try to achieve buy-in a second time. A careful, diligent approach will enable you to introduce AI in a way that gives your team – and the entire organization – comfort as you transform into a modernized AML compliance program that has the right tools in place to keep pace with the fast-evolving threat landscape.
Visit workfusion.com to learn about the benefits of scaling your anti-financial crime compliance teams with the help of our AI Digital Workers.