Compliance Chiefs Turn to AI Agents for AML Workflows as Regulatory Signals Shift

Compliance Chiefs Turn to AI Agents for AML Workflows as Regulatory Signals Shift

Castellum.AI's CEO says financial institutions are racing to understand how autonomous AI systems can handle anti-money laundering tasks—and what regulators will accept as adequate oversight.

Peter Piatetsky, cofounder and CEO of Castellum.AI, sat down with Fintech Business Weekly's Jason Mikula on January 21 to discuss what he called "one of the hottest topics in fintech, banking and compliance: agentic AI." The conversation centered on how compliance leaders should evaluate agent-based solutions and what regulatory signals mean for firms deploying autonomous systems in financial crime prevention.

The discussion comes as compliance departments face mounting pressure to process larger transaction volumes while maintaining regulatory standards. Agentic AI—systems that can act autonomously rather than simply providing recommendations—represents a potential shift in how banks and fintechs handle know-your-customer (KYC) and anti-money laundering (AML) workflows.

Piatetsky outlined how Castellum.AI's agents are currently powering AML and KYC workflows for financial institutions, though the podcast format provided limited specifics on deployment scale or client names. The conversation focused heavily on the practical challenges compliance officers face when considering these systems: how to evaluate vendors, how to document decision-making processes for examiners, and how to build model governance frameworks that satisfy regulators.

The regulatory dimension looms large. Piatetsky and Mikula discussed what regulators are signaling about AI agent adoption—a critical concern for CFOs whose institutions could face enforcement actions if oversight frameworks prove inadequate. The discussion touched on translating those regulatory signals into governance structures and examiner-ready controls, the documentation that becomes crucial during regulatory examinations.

Model governance emerged as a particular challenge. Traditional machine learning models typically have defined parameters and predictable outputs. Agentic systems, by contrast, can take actions and make decisions with less direct human oversight at each step. This creates thorny questions about accountability, auditability, and the level of human review required to satisfy both internal risk management and external regulatory expectations.

The conversation also looked forward, with Piatetsky offering views on how agents will reshape compliance programs over the coming years. For finance leaders, the implications extend beyond compliance: these systems could materially affect headcount planning, technology budgets, and the skill sets required in compliance departments.

The timing is notable. As of early 2026, financial institutions are moving past the "AI experimentation" phase and into deployment decisions that will affect their operational models for years. The question is no longer whether AI will play a role in compliance, but rather which types of AI systems will prove both effective and acceptable to regulators—and what the governance overhead will cost.

For CFOs evaluating vendor pitches and budget requests, Piatetsky's discussion with Mikula underscores a central tension: the technology may be ready, but the regulatory framework and internal governance structures are still taking shape. That creates both opportunity and risk, particularly for institutions that move too quickly without adequate controls or too slowly while competitors gain efficiency advantages.

The full interview, available on the Fintech Business Podcast, runs 52 minutes and 33 seconds—a length that itself suggests the complexity of deploying autonomous systems in one of banking's most scrutinized functions.