Nasdaq Verafin is developing two new agents: a fraud caseworker and a regulator reporter, adding to the company’s existing agentic workforce.

The agents to come, which have not been publicly announced, include a caseworker agent for investigative work and a regulatory reporting agent that would report anti-money laundering (AML) filings to regulators, Rob Norris, head of product strategy at the Canadian firm, which helps FIs protect themselves against financial crimes, told FinAi News.

Norris declined to comment further on the agents Nasdaq Verafin has not announced.

The latest agents

Other agents announced in June focused on alerting FIs about potential fraud:

• An agentic AML analyst, focused on cash structuring alerts; and
• An agentic fraud analyst focused on Automated Clearing House (ACH) fraud.

Cash structuring is when fraudsters break up transactions to avoid triggering the U.S. $10,000 currency transaction reporting threshold.

The fraud analyst targets ACH fraud in compliance with National Automated Clearing House Association regulations that took effect this spring, requiring more scrutiny of name matching in payments, Norris said.

Both agents, which will be available in Q3, run on a level one triage model, screening alerts, documenting rationale, closing likely false positives and escalating higher-risk cases to human investigators, he said.

“Our vision of the future is that a lot of the more repetitive and more simple activities can be automatically worked through the entire platform,” he said, adding that human investigators at FIs should focus on discovering sophisticated crime rings.

Industry-wide trend

Other tech providers are rolling out fraud prevention agents.

This spring, Anthropic and FIS collaborated to build an agent to fight financial crime, with early adopters including BMO and Amalgamated Bank.

Global Relay, a tech-driven compliance solutions provider for FIs, is also using agentic AI for risk detection, and expects to eventually deploy agents that produce data-driven insights, Donald McElligott, vice president of compliance supervision told FinAi News.

Monotonous risk evaluation tasks will eventually be automated by agentic AI, McElligott said.

“Everybody is going where the pain points are, which is too many false positives, too dull of a job,” McElligott said. “No reviewer is able to hold focus over the course of the day.”

From copilot to autopilot

By the end of the year, Nasdaq Verafin’s agentic AML and agentic fraud analysts will expand their capabilities to include auto-dispositioning — letting configured agents close or escalate alerts without a human in the loop.

That marks a shift from the agents’ previous recommendation-only mode. Banks can specify where they want autonomy to apply, according to Scott Peddle, head of AML product strategy at Nasdaq Verafin, in the June release.

“If a wire was for $1 million, the institution can decide, ‘You know what, I’m not going to let the analyst decide the million-dollar alerts,'” Peddle said.

Nasdaq Verafin is also working on a platform-agnostic product that lets the agents log into and work alerts inside third-party case management systems while still drawing on Verafin’s consortium data, per the release.

Beta testing on the platform-agnostic model and the new analysts will begin in the second half of 2026, Norris said.

Early results

A previously released agent designed by Verafin for sanction alerts delivered about 90% efficiency on alert queues at some banks, and the company’s EDD Analyst cut review time by about 50%, Norris said.

First National Bank of Omaha, a named customer, reached 100% agreement with the Sanctions Analyst’s recommendations in one review period, he said. Metrics on the AML and fraud analysts are not yet available, but Norris said he expects them to land in the same range.

The agentic workforce is model-agnostic and runs through AWS Bedrock, Norris said.

Register here for the FinAi Lending Summit, set for Oct. 7-8 in Las Vegas.