.svg)
Napier AI, the London-based financial crime compliance technology company, has received Honourable Mention in the Best AML Transaction Monitoring Innovation category of the Datos Insights 2026 Fraud Impact Awards for its Agentic Test Simulator for AML Effectiveness.
Currently in closed beta, the Agentic Test Simulator is designed to help financial institutions objectively assess how effectively their existing transaction monitoring systems detect money laundering. It stress-tests an institution’s rules and AI models against synthetic financial crime data generated by autonomous agents, identifying potential gaps across different typologies.
Financial institutions have traditionally assessed transaction monitoring performance through measures such as alert volumes, system tuning and retrospective validation. While these approaches provide useful operational insight, they do not necessarily establish whether a system would identify genuine criminal behaviour.
Napier AI’s Agentic Test Simulator instead evaluates detection effectiveness by representing each synthetic customer as an autonomous agent whose behaviour evolves over time. As the agents interact and coordinate, complex financial crime typologies including mule networks, smurfing and circular payments emerge dynamically, rather than being drawn from fixed transaction templates.
Because the simulation is controlled from end to end, each illicit event is known and labelled. This allows institutions to objectively measure recall and typology coverage, understand which activity their existing systems would detect and identify where potential vulnerabilities remain.
The capability incorporates learnings from Napier AI’s work with the UK Financial Conduct Authority (FCA), including the development and validation of money laundering typologies using large-scale synthetic financial crime data. Following each test, institutions receive a structured report covering overall detection coverage, performance by typology and identified gaps, supporting engagement with regulators, internal audit and model risk teams.
Sonnie Morse, Product Lead for Transaction Monitoring at Napier AI, commented: “Alert volumes can show how busy a transaction monitoring system is, but they cannot prove whether it would catch a money launderer. The Agentic Test Simulator enables compliance teams to see, typology by typology, where their detection performs effectively and where vulnerabilities may remain. “This recognition from Datos Insights reflects the importance of moving beyond traditional alert tuning towards objective evidence of detection effectiveness. By giving financial institutions a clearer and more defensible view of how their systems would respond to modern money laundering, we can help them strengthen their controls before those weaknesses are exploited by criminals.”
In its assessment, Datos Insights recognised the Agentic Test Simulator for addressing a longstanding gap in transaction monitoring validation. The analyst firm highlighted its use of autonomous agents to create realistic, evolving financial crime behaviour, alongside the regulatory grounding provided by Napier AI’s work with the FCA.
“Financial crime compliance functions often struggle with knowing how effective their transaction monitoring frameworks are at detecting modern money laundering,” says Chuck Subrt, Fraud & AML Practice Director, Datos Insights. “Napier AI’s Agentic Test Simulator and its use of autonomous agents generate realistic, emergent financial crime behavior, producing simulation conditions that enable institutions to evidence AML detection effectiveness, identify detection gaps, and strengthen transaction monitoring.”
Early internal and pre-release testing has demonstrated the simulator’s ability to assess detection performance across up to 30 money laundering typologies and generate fully labelled synthetic datasets containing hundreds of millions of transactions within hours. As the capability remains in closed beta, these findings are indicative of its potential rather than validated customer outcomes.
Over the next 12 to 18 months, Napier AI plans to develop the Agentic Test Simulator from a controlled, service-led capability into a self-service feature within the Napier AI Continuum platform. This will allow institutions to test their transaction monitoring controls, implement changes to rules or models and reassess detection performance iteratively.
Read why Datos Insights selected the Napier AI innovation, in this case study
