Interview with David Caruso, Vice President of Financial Crime Compliance, WorkFusion
- AI has become a buzzword across banking and fintech. From your perspective, where is AI having real, measurable impact in financial crime prevention today?
Understandably, some might think that AI is just hype. And in some cases, it is. But in Financial Crime Compliance, it is very real and making a measurable impact today. The main reason for this is that much of Financial Crime Compliance work is what I call “programmable.” Since these programs are highly regulated, there must be clear, step-by-step procedures for how to identify, investigate, and resolve potential money laundering and fraud. For example, when a financial institution’s monitoring systems screen all customers against sanctions lists issued by the U.S. Department of the Treasury or the UK Treasury, and a potential match is found, the procedures include analyzing whether it’s a true match. AI can automate those procedures, display records, and suggest a resolution for a person to decide. This saves over 80% of the manual work that was traditionally done before the decision-making step.
- WorkFusion’s AI Agents have gained attention in compliance circles. Can you share a quick example of one in action?
Our AI Agents have names, just like any coworker. One agent is Tara. Tara helps examine and resolve payment screening alerts. Processing payments quickly and accurately is a core banking function. But processing payments for individuals or businesses on a U.S. Treasury Office of Foreign Assets Control (OFAC) sanctions list is a violation of law. So, payment screening systems flag potential sanction violations, and banks must investigate and resolve them to either approve or block the transaction. A mistake in either case can lead to legal risks or an unhappy customer whose payment isn’t processed. The number of payment screening alerts can be very high, and having people investigate and resolve them can take too much time. Tara investigates and resolves these alerts in under a second, ensuring payments are processed quickly or immediately blocked if needed. - Is the pace of regulation keeping up with the speed of AI innovation or are we heading for friction?
Regulation is trying to keep pace, and that should be acknowledged. However, the speed of AI development is unique. This often slows down adoption because financial institutions want to ensure they aren’t moving too quickly, which could cause regulators to become overly anxious. Nonetheless, there are effective regulatory compliance frameworks that help banks evaluate, test, and adopt AI. For example, requiring financial institutions to have a written program detailing how they select or build AI-enabled solutions, how they assess their effectiveness, and whether the applications are explainable and operate consistently at scale. Regulators across the U.S., Europe, and Asia actively encourage financial institutions to explore AI usage. There is consensus that modern technology can improve banking, finance, and compliance, but like any new technology, it must be chosen, implemented, and operated in a clear and safe manner. - What’s one example that captures what’s possible when AI meets compliance? Something that says, “This is what the future looks like.”
An area that excites me is how AI is transforming suspicious activity reporting (SAR) and investigation processes. Investigations that lead to SARs are the most crucial part of Financial Crime Compliance. In fact, identifying, investigating, and reporting suspicious activity is the primary reason AML laws exist. For the past 30 years, these investigations required hours of work, much of it involving gathering and organizing information. For an eight-hour investigation, four or five hours were spent searching bank systems, the web, and copying and pasting records into files. AI now automates all of this, cutting investigation time by 50% or more. Building on that, AI can analyze activity, apply reasoning, and suggest outcomes to investigators. Soon, investigators will be able to multiply their capacity to complete investigations by five times or more. The importance of this impact on Financial Crime Compliance cannot be overstated.
- You’ve seen compliance evolve from spreadsheets to smart automation. What excites you in compliance right now?
The most exciting part is that AI has shifted from a point where Financial Crime Compliance executives view it skeptically to a situation where every leader in the field understands they must integrate it into their programs. We are now at a stage where not using AI is riskier than using it (though using it correctly is not risky).
- What advice would you give compliance teams trying to build AI into their operations?
Begin with payment screening, name screening, or adverse media monitoring. These are all areas where AI can make a quick and meaningful impact. AI has proven effective over the past few years in handling this work millions of times. It’s a great way to see the benefits of AI and the savings it generates. Use those savings to fund the next phase of AI, which is suspicious activity investigation and reporting.
- If you had 30 seconds to pitch AI to a skeptical customer, what would you say?
Financial Crime Compliance has largely remained unchanged for 20 years — manual, repetitive, and prone to error. The focus of workers and regulators is on strict adherence to procedures. This diverts attention from the real goal, which is identifying those who misuse our financial system to harm others. AI finally enables a transformation. It eliminates the tedious, repetitive, and manual tasks. Compliance with procedures becomes automatic. This allows workers to focus on what AML laws are meant for: detecting money launderers, financial terrorists, and fraudsters who threaten our citizens.
- Looking ahead to 2026, what trends do you think will help define the next era of financial crime compliance?
2025 was the year AI Agent adoption for Financial Crime Compliance became fully accepted. In 2026, I see a few things coming. One is that AI will establish new standards in Financial Crime Compliance. Just as earlier innovations, such as screening systems (30 years ago) and transaction monitoring systems (20 years ago), set standards that all financial institutions had to follow, AI will do the same. For example, continuous Adverse Media Monitoring where AI scours every news item published every day, all over the world, and identifies any mention of a bank’s customer. Currently, fewer than a handful of banks do this, but by year’s end, dozens will, and soon after, it will become widespread.
