Castellum AI Agent Passes Compliance Exam as Regulators Embrace AI

Key Highlights:

• Castellum.AI’s artificial intelligence (AI) agent passed the Certified Anti-Money Laundering Specialist (CAMS) practice exam on its first attempt, a feat typically reserved for human compliance officers.
• The New York-based startup closed an $8.5M Series A round led by Curql, with participation from BTech Consortium and Framework Venture Partners.
• The funding will accelerate product development and expand integrations for mid-tier financial institutions struggling to combat sophisticated fraud.

AI vs. Financial Crime: A New Arms Race

CEO Peter Piatetsky, a former U.S. Treasury sanctions officer, founded Castellum in 2019 with a focus on first-party data. Unlike competitors relying on third-party feeds, Castellum scrapes and standardizes data directly from hundreds of global government and press sources, updating every five minutes.

• AI agents leverage this proprietary database to screen sanctions, detect fraud, and draft suspicious activity reports (SARs) in milliseconds.
• The company corrects errors in government sanctions lists, maintaining a public “Department of Corrections” for regulators like OFAC and the UN.

Real-Time Compliance for Crypto and Traditional Finance

With real-time payments and crypto transactions demanding faster compliance checks, Castellum’s AI eliminates batch processing delays:

• Crypto exchanges and stablecoin issuers use its platform for instant sanctions screening.
• Community banks and credit unions can bring compliance back in-house, reducing reliance on offshore review teams.

Regulatory Shift: From Skepticism to Acceptance

Piatetsky noted regulators have moved from questioning AI to demanding proof it works. Castellum addresses this by:

• Training custom AI models for each client, tested on both known and unknown alerts.
• Delivering 90% fewer false positives by cleaning and enriching raw government data.

What’s Next?

Castellum plans to:

• Hire engineers for deeper core banking integrations.
• Have its AI agent officially certified by taking the proctored CAMS exam.
• Scale its platform to more institutions, emphasizing speed and accuracy in the fight against AI-powered financial crime.

“Our product works, and clients love it. It’s time to scale,” Piatetsky said.