Castellum.AI, a financial crime compliance platform featuring in-house risk data, AML/KYC screening, and AI agents, has announced the close of an oversubscribed $8.5 million Series A funding round. The company stated that its platform aims to eliminate the traditional tradeoff between compliance accuracy and risk mitigation.
The investment round was led by Curql, a fund supported by over 130 credit unions, including Navy Federal. Additional participation came from BTech Consortium, a fund backed by more than a dozen banks such as Customers Bank, and Framework Venture Partners, which receives backing from Tier 1 Canadian financial institutions including RBC. Existing investors Spider Capital, Remarkable Ventures, and Cameron Ventures also contributed to the round.
Peter Piatetsky, CEO and Co-Founder of Castellum.AI, commented that the funding from credit unions and banks validates the platform’s market alignment and its role in helping financial services identify and prevent financial crime. Nick Evans, President and CEO of Curql, highlighted the challenges financial institutions face from AI-enabled criminal activity and inefficient legacy systems, emphasizing the importance of Castellum.AI’s fast, precise, and explainable AI agents for next-generation technology needs.
Castellum.AI reports a 94% reduction in AML/KYC false positives and an 83% decrease in compliance review time without prior tuning. This efficiency has positioned the company as a compliance platform for major financial institutions, credit unions, sponsor banks, fintechs, and leading crypto exchanges seeking to streamline operations. Mike Sekits, Co-Founder and Managing Director of BTech Consortium, noted that Castellum.AI uniquely combines in-house data, screening, and AI agents, underscoring the importance of underlying data quality for effective AI in compliance.
Castellum.AI’s AI agents have successfully passed a CAMS practice exam on their first attempt. The AI agent training was guided by former regulators and incorporated AML/KYC best practices and regulatory guidance from bodies such as the OCC, NCUA, NY DFS, and California DFPI. The company is seeking official CAMS certification for its AI agent. This capability, combined with detailed audit records, aims to provide institutions with confidence in deploying AI agents for financial crime prevention.
To facilitate adoption, Castellum.AI works with clients on modular implementations that integrate with existing tech stacks and incorporate specific risk requirements, thereby reducing switching costs. Ajay Gopal, a partner at Framework Venture Partners, stated that this flexibility was a key factor in their investment decision, enabling rapid client support.
Castellum.AI employs an integrated compliance approach, consolidating risk data (including sanctions, PEPs, and adverse media), AML/KYC screening, and AI agents into a single, audit-ready system built internally. Piatetsky explained that the platform eliminates the need to delay transactions or onboarding to identify true positives, with its explainable AI supporting Level 1 and Level 2 alert adjudications and Level 3 investigations, all with detailed audit trails.
Client outcomes include an 88% reduction in AML/KYC review costs for a leading sponsor bank, a 94% reduction in transaction screening false positive rates for a community bank, an 83% reduction in Level 1 KYC alerts for a Fortune 50 corporation, and a reduction in data refresh time from 24 hours to every 5 minutes for a major crypto brokerage.
Castellum.AI provides clients with direct oversight of their compliance processes. Its patented data collection and machine learning enrichment process gathers risk data from original sources, feeding into its screening and monitoring system for real-time risk exposure notifications. Piatetsky highlighted features such as AI summaries for global adverse media, AI-powered enrichments for unstructured data categorization, and AI agents for alert review.
