- Covalent has made a growth investment in AML Analytics (AMLA)
- AI is driving growth. AI and agentic finance are reshaping how financial crime is committed, detected and regulated, and every one of those shifts raises the need for independent testing
- AMLA is an AI Thriver. Data gravity and mission criticality are what make a company stronger as AI advances, and AMLA has both
- Justin Holder founded AMLA in 2010 and built it into the global standard, trusted by more than 50 regulators and 1,450 institutions. Covalent is partnering with Justin and the team to accelerate the engineering roadmap, deploy AI-native testing frameworks, and expand global distribution
// What AMLA Does
AML Analytics is the global category leader in testing and measuring the effectiveness of financial crime systems for financial institutions, payment platforms, and regulators. Its method has four steps, and the fourth is what makes it a benchmark rather than a tool.
- Build test records
- Run them through the live system
- Score the result
- Compare with peers
The same test gives the same result, so a compliance team and its examiner read the same score.
// AI Is Driving Growth
Financial crime is moving at machine speed. AI is empowering bad actors to launch sophisticated attacks to evade sanctions and launder money. On the other side of the screen, banks and payment platforms are handing their controls to machine learning models that change with every retraining. These shifts create a growing need for independent benchmarks to measure the efficacy of defense systems that change continuously.
On the attack side, FinCEN warned banks in November 2024 that deepfake identity documents were rising in suspicious activity reports, and Regula found 49% of businesses met video deepfake fraud in 2024, up from 29% in 2022. On the defense side, banks are moving from fixed rules to machine learning. Adoption is early: a 2025 Chartis survey found only about a fifth of banks run AI at scale in transaction monitoring, while EY found 75% of Nordic banks plan to invest further. These shifts create demand for reference data and independent benchmarks to tune and validate dynamic and generative compliance systems.
// Offense: criminals with AI
- Synthetic identities from generated faces and documents.
- Name mutation at scale to find the variants a filter misses.
- Automated probing of thresholds before the real payment is sent.
// Defense: banks with AI
- ML monitoring replaces fixed rules: fewer alerts, less transparency.
- Frequent retraining. Each version changes what the system catches.
- Guidance gap. April 2026 US guidance withdrew the AML model statement; generative AI is out of scope.
// The AI Thriver Matrix
At Covalent we invest in companies that AI makes stronger. Our AI Thriver Matrix scores a company on two axes. Data Gravity is how much unique data it holds and whether that data grows with use. Workflow Surface Area is how much mission-critical work runs through the product, and so where AI agents will act. AMLA scores high on both.