How is AI Transforming Banking Compliance?

Fidelity National Information Services (FIS) and Anthropic announced a partnership this week to deploy agentic AI systems inside banking operations, beginning with anti-money-laundering (AML) and financial crimes investigations.

The announcement was made at FIS Emerald in Orlando. According to FIS, BMO and Amalgamated Bank will be among the first institutions deploying the new Financial Crimes AI Agent, with broader availability expected in the second half of 2026. The company said the system is designed to reduce investigation workflows from hours to minutes by assembling evidence across banking systems and surfacing higher-risk cases for review.

FIS said U.S. financial institutions spend between $35 billion and $40 billion annually on AML operations, much of it tied to manual compliance work and fragmented systems. Early testing reduced manual process work by as much as 90%, according to the company.

The launch adds to a broader shift across banking, where AI deployments are increasingly moving beyond customer-facing copilots into regulated operational systems tied to compliance, fraud monitoring, and transaction infrastructure. Similar efforts are emerging across financial services, including AI systems built for trade surveillance, fraud review, and governed banking workflows.

FIS said Anthropic’s Applied AI and forward-deployed engineering teams worked directly with FIS to build the platform. Customer data remains inside FIS-controlled infrastructure, according to the company.

Stephanie Ferris, CEO and President of FIS, said the company’s regulatory and compliance experience is central to its AI strategy.

“Our moat is the 50 years of deep regulatory and compliance experience,” Ferris told Forbes in an interview published on May 6.

Jonathan Pelosi, Head of Financial Services at Anthropic, said the partnership combines Anthropic’s reasoning models with FIS’ banking infrastructure and regulatory expertise.

Compliance and Governance Become Core AI Infrastructure

The partnership reflects a larger trend inside enterprise banking AI, where governance and orchestration layers are becoming as important as the underlying models.

Banks have historically moved slowly on AI deployments because regulated systems require explainability, auditability, traceability, and human oversight. Financial crimes investigations represent one of the most tightly controlled operational environments in banking, making AML a critical testing ground for autonomous systems.

FIS’ strategy centers on controlling the infrastructure layer surrounding AI deployment rather than building its own frontier model. The company said it maintains governance, compliance architecture, deployment systems, and banking relationships while Anthropic provides reasoning capabilities through Claude.

That positioning mirrors a broader industry push toward controlled AI execution environments inside financial institutions.

The announcement also follows FIS’ earlier “agentic commerce” push after its acquisition of Global Payments’ Issuer Solutions business. At the time, Ferris argued banks needed infrastructure capable of managing AI-initiated transactions safely inside regulated payment systems.

Payments and banking companies are increasingly investing in AI systems tied to fraud detection, transaction intelligence, and operational automation rather than standalone chat experiences.

Anthropic separately announced a broader expansion into financial services this week, including AI agents designed for institutional finance workflows such as earnings review, credit memo drafting, and financial statement analysis.