New AML Agent Cuts Bank Investigation Time by 90 Percent

The AI-driven tool automates data assembly and evidence gathering to address industry-wide false-positive bottlenecks.

Updated on Oct. 15, 2026 in Artificial Intelligence

Isometric editorial illustration featuring a vertical stack of clean, organized translucent blocks representing data, illustrating banking compliance automation. — FIS has launched an AI-driven investigative agent for U.S. banks designed to reduce false-positive rates in money-laundering detection by 90 percent. AI Illustration. Upload story photo >

Live Poll

Do you trust artificial intelligence to handle the complex investigation of financial crimes in banking?

FIS has launched the AML Investigations Agent for U.S. banks, an AI tool built on Anthropic Claude models that automates data assembly for financial crime checks. The system is now commercially available to all U.S. banks, regardless of their current core banking provider.

Why it matters

Financial institutions currently face 90 to 95 percent false-positive rates on money-laundering alerts, consuming resources that could otherwise be dedicated to complex case analysis. This deployment aims to shift the focus of human investigators away from repetitive data collection toward high-level decision-making.

The platform is projected to reduce Level 1 alert review time by 90 percent, Level 2 investigation time by 98 percent, and Suspicious Activity Report (SAR) narrative generation time by 80 percent. These gains are measured against baseline industry averages where analysts spend 40 to 60 percent of their time manually assembling case data.

The players

FIS

A provider of financial technology solutions, including the Horizon and Systematics core banking platforms.

Anthropic

An AI research and deployment company known for its Claude family of large language models.

Amalgamated Bank

A financial institution that served as a design partner for the development of the AML Investigations Agent.

The details

The agent operates in a private cloud environment, integrating with FIS core banking platforms to map customer types and historical activity against triggered alerts. The FIS Control Plane technology acts as a logical barrier, isolating each bank's specific risk thresholds and internal procedures from the broader system. Each AI-generated output is linked to specific records, with all recommendations and findings stored in a formal audit trail to ensure compliance transparency.

2023: Reported industry costs and alert false-positive rates.
2025: Reported percentage of investigation time dedicated to data assembly.
May 2026: FIS and Anthropic announced the development of the AML agent.
October 15, 2026: The AML Investigations Agent became commercially available.

The Tech Race

This release positions FIS against a growing movement of fintech firms using generative AI to solve the high-cost, high-friction problem of AML compliance. It marks a shift from experimental LLM integration toward enterprise-grade, audit-ready banking tools that prioritize traceable decision logic.

The agent is available to all U.S. banks, including institutions not currently utilizing FIS infrastructure. Financial investigators can expect to transition from manual spreadsheet and document synthesis to a system that automatically links findings to source records.

The takeaway

The deployment of this agent signals a transition from AI research to the active reduction of manual labor in high-stakes financial compliance. Banks should track the adoption rate of these audit-ready agents to see if the projected 90 percent reduction in review time manifests in real-world quarterly filings.