Financial crime is growing faster than the systems built to catch it. An estimated $4.4 trillion in illicit funds moved through the global financial system in 2025, a nearly 20% compound annual growth rate since 2023, according to Nasdaq Verafin's 2026 Global Financial Crime Report.
Compliance teams, meanwhile, are stretched thin: nearly 60% of anti-financial crime professionals cite inadequate resources, human and technological, as a top concern.
Criminal operations increasingly use AI themselves to scale their operations, without facing the model risk committees, governance reviews, or regulatory exams that slow down the institutions trying to stop them.
Regulators, including the Financial Crimes Enforcement Network (FinCEN), the U.S. Treasury bureau responsible for combating money laundering and other financial crimes, have begun signaling openness to AI experimentation, treating innovation as part of the solution rather than a compliance risk to be managed around, as Nasdaq Verafin explored in a recent blog post on agentic AI in financial crime investigations.
First National Bank of Omaha (FNBO), a $34.6 billion-asset bank with 170 years of history, offers an early look at what that shift looks like in practice. After deploying two agentic AI analysts from Nasdaq Verafin to handle sanctions and enhanced due diligence (EDD) reviews, the bank cut the time investigators spend on those cases by 50%, according to a customer spotlight published by Nasdaq Verafin.
TL;DR
-
Global financial crime is scaling faster than compliance teams can keep up with, with an estimated $4.4 trillion in illicit funds moving through the system in 2025.
-
FNBO cut time spent on sanctions and EDD reviews by 50% after deploying agentic AI analysts from Nasdaq Verafin, redirecting up to 50% more investigator capacity to deeper analysis.
-
The bigger story isn't the tool itself, it's how quickly regulated industries are being pushed to adopt autonomous systems they still have to govern and explain.
Why Are Financial Institutions Under So Much Pressure?
Most banks aren't short on effort. They're short on capacity relative to the scale of the problem. Manual investigation work, gathering data across systems, applying consistent logic case by case, documenting the reasoning, takes time that compliance teams don't have in a market where criminal activity is compounding at nearly 20% a year.
That gap is why agentic AI, systems that can reason through multi-step tasks and act with minimal human input, has moved from a research topic to a live deployment question for banks this size. It's not a hypothetical efficiency gain anymore. It's a response to a resourcing problem that isn't going away on its own.
What Is Agentic AI?
A category of artificial intelligence that operates autonomously toward a defined goal, reasoning through multi-step tasks, gathering its own supporting data, and documenting its logic, rather than waiting for step-by-step human prompts.
What Did FNBO Actually Change?
Before deployment, FNBO's investigators faced a familiar bottleneck: manual information gathering across multiple systems, taking a few minutes per sanctions alert and up to 20 minutes per EDD case, with no consistent starting point depending on who picked up the alert.
According to David Dawson, the bank's BSA/AML Officer (Bank Secrecy Act / Anti-Money Laundering), that inconsistency was itself a risk, not just a slowdown. After deploying Nasdaq Verafin's Agentic Sanctions Analyst and Agentic EDD Analyst, every case now starts with standardized, AI-generated documentation already in place, regardless of workload or which investigator picks it up. The AI gathers information and drafts documentation aligned to regulatory expectations; investigators still make the judgment calls, interpreting patterns, assessing risk, and deciding whether to escalate or clear.
"Nasdaq Verafin's Agentic AI Analysts are having an immediate impact, reducing our time spent by 50% compared to before deployment. We're also seeing up to 50% more investigator capacity redirected to holistic analysis and complete, auditable records automatically generated for regulatory readiness." — David Dawson, BSA/AML Officer, FNBO
The bank ran parallel testing on real cases while its Quality Assurance (QA) team evaluated every AI output against existing investigative standards, only scaling up once that validation held. In that testing, Dawson said the AI identified 100% of cases requiring further review.
"The quality improvements we've seen are huge. Documentation is consistent across every case, which strengthens our regulatory position. And in our quality assurance, the Agentic Analysts also identify 100% of cases requiring further review. We're not just working differently, we're working smarter."
Does This Replace Human Judgment?
The bank operates on a human-in-the-loop model: the AI issues a recommendation, but an investigator retains final authority. No alert is cleared, and no case is escalated without that review. Nick Baxter, FNBO's Chief Risk Officer, described the decision less as a leap into new territory and more as an extension of an existing relationship with its vendor:
"Onboarding agentic AI wasn't a major leap for us. Nasdaq Verafin is a true partner, and I don't use that term lightly. They have decades of AI expertise, and their solutions have been embedded in how we operate for years." — Nick Baxter, Chief Risk Officer, FNBO
That framing matters for how other institutions might approach this. The technology itself is only part of the equation; the pace of adoption seems to hinge as much on institutional trust and existing vendor relationships as on the tool's capabilities.
