Why AI competency is the new baseline for compliance: Key takeaways from our New York City roundtable

Our team recently had the pleasure of hosting discussions with experts from New York. Participants in our discussions included senior financial crime leaders, with the goal being to pressure-test the findings of our State of Financial Crime 2026 report. While our global data provides a high-level map of the industry, there is no substitute for discussing industry trends with the people navigating risk and regulatory shifts in real-time.

What struck me most during our conversation was how quickly the future has become today’s baseline. In the United States, we are moving past the era of speculating about AI and into a phase of comprehensive, practical use and deployment. From the talent we hire to the way we calibrate friction in the customer journey, the expectations for compliance and financial crime risk management functions have fundamentally transformed.

Here are three key takeaways from our discussion that I believe every leader in this space should consider.

1. The elevation of technical competency

A significant theme of the discussion was the changing profile of the modern compliance professional. In leading firms, AI competency is no longer viewed as a specialized skill set but as a core competency.

We are seeing a shift in which Boards expect Chief Compliance Officers to have the mandate and tools to integrate AI across their functions. Participants noted that the bar for entry has risen so rapidly that continuous technical upskilling is now a survival strategy for compliance teams. To lead effectively in 2026, understanding the underlying logic of your AI stack is just as important as understanding the regulations themselves.

2. Strategic friction vs. “minimum viable compliance”

We explored the tension between rapid growth and regulatory stability, particularly within the FinTech sector. While the pressure to launch quickly can lead to “minimum viable compliance” models, the consensus among the leaders in the discussions was that this approach often leads to significant “technical debt.”

The goal for a mature financial crime risk management function is to apply the necessary and sufficient amount of friction required to protect the business without impeding its trajectory. Modernization is not about removing safeguards, but about making them smarter and more integrated so that they scale naturally with the business.

3. Closing the execution gap with agentic AI

Our 2026 survey revealed a striking disconnect: 100% of organizations have achieved or expect to see positive outcomes from adopting agentic and predictive AI. The industry is fully aligned on the potential value, yet this optimism hasn’t yet been reflected in adoption rates. Currently, only 33% of firms are using agentic solutions for customer screening, and just 32% for transaction monitoring.

In New York, we discussed why this “execution gap” persists. While the intent is there, many firms are still held back by manual bottlenecks; in fact our data shows 81% of North American organizations take over five minutes to clear a single sanctions alert.

The consensus among the leaders that we spoke with was that moving from expectation to adoption requires two things: better detection and unwavering explainability. For US firms, AI cannot be a black box. To move the needle on those adoption stats, agentic systems must be able to provide a clear, human-readable rationale for every action and decision. Only then can we optimize and enhance manual operations with the real-time, co-pilot defenses that the market is clearly hungry for.