Alviere, a company that offers a variety of financial products, has been testing whether agentic AI can perform anti-money laundering investigations to the same standard as its human analysts.
The firm ran Hawk’s AML Investigative Agent against a batch of historical alerts, measuring the tool’s conclusions against decisions its own analysts had already made on identical cases. The outcome showed the agent escalating every alert that had been confirmed as a genuine true positive, meaning no real risk slipped through unnoticed. It also recommended closure for 98% of the alerts analysts had previously ruled to be false positives, while independently surfacing a number of additional cases that appeared genuinely suspicious.
Because Alviere serves such a varied customer base, behaviour that would raise concern in one part of the business can look entirely routine in another. Its team had been leaning on manual adjustments to detection rules to account for this, but constantly recalibrating thresholds across so many different use cases put a strain on resources and pulled analysts away from the more complex reviews that needed their attention most.
To ease this burden, Alviere rolled out the Hawk Investigative Agent across four separate business lines, building the workflow around its existing AML policies so the tool mirrored the firm’s own investigative approach. For every alert, the agent followed the same sequence: reviewing the alert itself, pulling in customer information, checking whether the source of funds and account activity matched the customer’s profile, and tracing counterparties as far as necessary.
Business-specific context was layered into the red-flag stage of each review, giving the agent a working sense of what counted as ordinary activity within each line of business. The system also used network graph analysis to map how funds moved between accounts, helping analysts spot patterns such as circular transfers, hub-and-spoke structures, or shared addresses and employers linking otherwise separate customers.
Alviere had already seen efficiency improvements after adopting Hawk’s AML Transaction Monitoring solution.
Over three months of alerts spanning the four business lines, the agent uncovered several cases that illustrated its ability to catch risks a narrower review might miss. In one instance, a sudden 400-times increase in transfer size led the agent to a counterparty account acting as a fan-in/fan-out hub, drawing funds from multiple sources before rapidly redistributing them to unrelated recipients overseas.
In another, an account that appeared unremarkable on its own was found to be transacting with a counterparty involved in same-second round-trip transfers and a prior AML history. A third case began with a modest $500 transfer that traced back to a cash-fed hub feeding a wider layering network, with funds fanned out to Mexico and to a broker with no apparent connection to the original account holder.
For more, read the full case study here.
Copyright © 2026 FinTech Global
