FIS saw its stock rise about 5% following news that the company has formed a partnership with Anthropic to develop an AI-driven assistant focused on financial crimes investigations for banks.
The Financial Crimes AI Agent is designed to speed investigations by automatically collecting and assembling evidence from a bank's core systems and by assessing activity against recognized patterns of illicit conduct. The stated aim is to reduce investigation timelines that currently take hours down to minutes, thereby shortening the period before substantive analysis can begin.
According to FIS, BMO and Amalgamated Bank are set to participate as early testers of the technology. The company said it expects wider availability of the agent in the second half of 2026.
Anthropic's Applied AI team, together with forward-deployed engineers, is working directly with FIS on the agent's design. The tool leverages Anthropic's Claude models for reasoning functions while operating within infrastructure controlled by FIS, a configuration FIS says is intended to preserve data security and meet regulatory requirements.
The announcement cited estimates that underscore the scale of the problem the tool targets: roughly $2 trillion in illicit funds move through the global financial system each year, and U.S. financial institutions spend between $35 billion and $40 billion annually on anti-money-laundering operations. Investigators commonly spend substantial time on manual evidence collection prior to analysis, a step the agent is designed to automate.
FIS indicated the agent's deployment should lower the cost per case by removing the need for manual evidence gathering and by cutting down case review durations. The company also emphasized that human investigators will retain authority over final decisions; the agent's role is to handle the initial assembly and analysis of relevant data.
Beyond financial-crimes investigations, FIS plans to broaden the agent platform to address other banking functions. The company listed potential future applications including credit decisioning, deposit retention, customer onboarding and fraud prevention. FIS serves as the system of record for transactions and customer data for thousands of financial institutions, positioning it to integrate the agent across a broad customer base.
Implications and context
The partnership pairs Anthropic's reasoning models with FIS's institutional infrastructure and customer footprint. If the agent functions as described, it could alter workflow economics in compliance and fraud teams by shifting time spent on data collection to automated processes, while leaving adjudication and final determinations to human investigators.