Feedzai Survey Finds AI Is Delivering 40%+ Productivity Gains for Most AML Teams

Most anti-money laundering (AML) teams that use AI are reporting productivity gains above 40%, according to new research from Feedzai, as banks face rising transaction volumes, increasingly sophisticated financial crime, and tighter regulatory scrutiny. Feedzai said the findings show AI is assisting institutions “cut through alert overload,” “reduce false positives,” and prove that controls are delivering “real-world risk reduction, not just passing audits.”

The results are published in The AI Shift: Transforming AML Compliance into Competitive Advantage, which Feedzai describes as being based on its 2025 surveys of “global AML and financial crime professionals.” The report positions the industry at what it calls “a clear inflection point,” as supervisors move from reviewing whether processes exist to assessing whether those processes actually work in practice.

Feedzai’s framing is that this shift is not merely operational but strategic: teams that can demonstrate effectiveness while handling growing volumes are positioned to turn compliance into performance. In the company’s words, “AI is assisting institutions” move away from being swamped by activity metrics and toward controls that can be explained, audited, and shown to reduce risk outcomes.

Regulators Want Proof of Effectiveness, Not Just “Alerts Reviewed” and “SARs Filed”

Feedzai says the compliance goalposts are moving. “For years, AML success was measured by activity: alerts reviewed, rules triggered, SARs filed.” But the firm says regulators are now asking for evidence that controls are “reducing illicit financial flows” and that AML resources are directed “where risk is highest,” rather than simply documenting high volumes of investigative activity.

Nuno Sebastião, CEO of Feedzai, tied the shift directly to supervisory expectations, saying: “For years, AML programs were measured by how much they processed.” He added: “Now regulators want proof that controls actually reduce risk.” In his view, AI supports that transition because “AI assists volume metrics, focus analysts where it matters most, and show real impact without compromising transparency or trust.”

In practical terms, the report argues that outcome-driven oversight depends on models that compliance teams can defend. Feedzai’s survey data underscores that point: “Trust is non-negotiable,” it says, with “95%” of respondents stating that “AI must be explainable and auditable,” reinforcing that banks need governance as much as they need automation.

Efficiency Gains, False-Positive Reductions, and Explainability Shape the Competitive Edge

The “clear, quantifiable advantages” for teams using AI to prioritise risk. Among AML teams using AI, “66%” “productivity improvements above 40%,” which Feedzai says assists ease “alert backlogs and analyst strain.” The company also points to quality improvements, noting “62%” report “false-positive reductions above 40%,” which it says allows teams to focus on “higher-risk activity” rather than churn through noise.

Feedzai’s data also suggests the regulatory climate is becoming more supportive of AI—provided institutions can show robust controls. The survey found “96%” of AML professionals say regulators now “encourage AI adoption,” a signal—according to Feedzai—that supervision is increasingly aligned with modernisation as long as outcomes, governance, and transparency remain intact.

Feedzai’s broader message is that this is “an evolution, not a teardown.” The company argues banks do not need to dismantle existing AML programmes; instead, they can “modernize responsibly,” strengthening detection while maintaining “governance, transparency, and human oversight.” In Feedzai’s framing, the institutions that adapt now will be better positioned to meet rising expectations, reduce operational strain, and make effectiveness “a requirement,” while turning “operational excellence” into a “competitive advantage.”

“`