Castellum.AI

What Castellum.AI does

Castellum.AI helps AML teams eliminate alert overload with AI agents that resolve alerts across customer onboarding, payments, and investigation workflows. These agents enable teams to process investigations 6x faster while eliminating 94% of false positives. Arbiter agents are trained on each institution's internal policies and procedures, ensuring decisions reflect specific risk appetite. They can be deployed as a modular AI layer within existing workflows or as part of a consolidated Castellum.AI solution that includes global risk data (sanctions, PEPs, adverse media, etc.) with real-time screening and monitoring. The company aims to empower compliance teams to focus on preventing financial crime.

Products

• Arbiter: AI Agents for FinCrimeResolves L1 and L2 alerts for onboarding, transaction screening, pKYC, and more in less than a second. These AI agents are policy-trained and provide explainable, audit-ready documentation.
• KYC/KYB OnboardingInstantly identifies risky customers and streamlines the onboarding process for others.
• AI-Powered Adverse MediaProvides exceptionally reliable results that lead to higher value alerts and faster alert reviews, with real-time updates and AI summaries.
• Transaction ScreeningScreens large transaction volumes in real time against global risk data with 94% fewer false positives.
• Enriched Sanctions and PEPs dataDetects real threats by screening 1,000+ watchlists updated every 5 minutes, minimizing false positives from similar names.
• Real-time MonitoringEnsures customers never see the same alert twice and receive instant notice of changes in a customer’s risk profile.
• Sift: Transaction Monitoring AgentsAdaptive AI agents that flag and adjudicate suspicious activity in seconds, continuously learning from emerging risks and fraud typologies.
• API IntegrationDeveloper-friendly API for integrating Castellum.AI into compliance workflows, offering global risk screening and Agentic AI alert adjudication.

Key capabilities

• Agentic Screening and Alert Resolution
• 94% reduction in alert volume
• 83% less time on L1 alert reviews
• 5-minute data update cycle
• Global Data Coverage (200,000+ sources across sanctions, PEPs, adverse media)
• Real Time AML/KYC Screening
• Explainable AI Agents
• Modular implementation
• Audit-ready documentation
• Customizable risk controls
• Built for highly-regulated workflows
• Enterprise-grade scalability
• SOC 2 compliant
• Patented Jgram search algorithm for false positive reduction
• Automated Data Enrichment
• High Customization Controls
• Unified Data Structure
• 100+ Language Support
• Real-Time Data Updates (every 5 seconds for adverse media, every 5 minutes for watchlists)
• AI-generated summaries for adverse media
• 99% Alert reduction for adverse media
• 20+ years of historical adverse media data
• Victim-perpetrator-bystander analysis for adverse media
• Integrated Screening and Monitoring for crypto
• Push Notifications for Crypto Sanctions
• Global Crypto Coverage
• Consolidated Crypto Sanctions Report
• Adaptive Risk Detection (for Sift agents)
• Built-In Compliance Guardrails (for Sift agents)
• Global Typology Coverage (for Sift agents)
• Real-Time Fraud and AML Defense (for Sift agents)

Use cases

• AML/KYC alert resolution
• Customer onboarding
• Payments screening
• Investigation workflows
• Transaction screening
• pKYC (periodic Know Your Customer)
• Third Party Risk Management (TPRM)
• Supply Chain Compliance
• Crypto compliance (screening transactions for sanctions, illicit activity, high-risk counterparties)
• Real-time monitoring of addresses, wallets, and customers
• Detecting cross-border risks (ACH, FedNow, Zelle, checks)

AI approach

Castellum.AI develops proprietary AI agents, called Arbiter and Sift, that are trained on internal policies and regulatory guidance to automate AML/KYC alert resolution, transaction monitoring, and adverse media analysis. These agents are designed to reduce false positives, speed up alert review, and provide audit-ready documentation. The company emphasizes explainable AI and customizable risk controls.

Tech named: AI agents, Jgram search algorithm, Explainable AI

Industries served

• Financial Services
• Banks
• Credit Unions
• Fintechs
• Digital Asset Firms
• Highly-regulated spaces

What it says sets it apart

• AI agents trained on each institution's own internal policies and procedures
• Eliminates 94% of false positives
• Resolves alerts 6x faster
• Modular AI layer that integrates with existing workflows
• Proprietary and enriched global risk data from 200,000+ sources
• Real-time data updates (every 5 minutes for watchlists, every 5 seconds for adverse media)
• Explainable, audit-ready, policy-trained AI agents
• Guarantees false positive reductions
• Built in-house AI agents and data
• Regulator-aligned AI (trained on guidance from OCC, NCUA, NY DFS, California DFPI, OFAC, FinCEN)
• SOC 2 Type II certified
• Patented Jgram search algorithm
• 100+ language support for screening
• Adaptive AI agents that continuously learn from emerging risks and fraud typologies (Sift)
• No rip-and-replace required for integration

Funding rounds we track

From the AI funding tracker — rounds as reported by the linked publications.