FIS has partnered with Anthropic to deploy an agentic AI system targeting anti-money laundering operations at financial institutions.
The agent will automatically assemble evidence across a bank's core systems, evaluate activity against known typologies, and surface high-risk cases for investigator review. BMO and Amalgamated Bank are currently in development with the tool, with broader availability to FIS clients planned for the second half of 2026.
From manual workflows to automated investigations
Anti-money laundering operations remain one of the most resource-intensive functions within financial institutions. According to figures cited in the announcement, financial institutions in the US alone spend between USD 35 billion and USD 40 billion annually on AML operations, while the United Nations estimates that USD 2 trillion in illicit funds moves through the global financial system each year. Despite this scale, investigators currently spend the majority of their time manually gathering evidence across disconnected systems before substantive analysis can begin.
The Financial Crimes AI Agent is designed to address this directly. At case opening, the agent assembles the complete evidence package by connecting to a bank's core systems, whether operated by FIS or the institution itself, and evaluates findings against known typologies. Suspected activity reports and case narratives are also expected to improve in quality. The stated outcome is a compression of AML case and alert investigations from days to minutes, with investigators retaining decision authority throughout.
In addition, Anthropic's Applied AI team and forward-deployed engineers are embedded within FIS to co-design the agent and establish evaluation frameworks, with the stated intent of enabling FIS to build and scale additional agents independently over time.
Infrastructure, governance, and the agent roadmap
The deployment is built around an architecture in which client data remains within FIS-controlled infrastructure at all times. Every agent decision is designed to be traceable and auditable, a requirement in regulated financial services environments where explainability and accountability are both operational and compliance considerations. For institutions not running FIS core systems, the agent connects via open integration standards, while the governance and audit layer remains within FIS infrastructure regardless of the source data's origin.
FIS positions itself as the data and infrastructure layer in this model, holding the system of record for transactions, payments, deposits, credit, and customer activity across thousands of financial institutions. Anthropic's Claude models provide the reasoning capabilities across the platform.
Moreover, financial crimes is described as the first proof point. The broader agent roadmap spans credit decisioning, deposit retention, customer onboarding, and fraud prevention, each to be delivered through the same governed platform. The collaboration reflects a wider industry trend of financial technology providers moving from AI-assisted tools toward AI agents capable of executing multi-step workflows autonomously within regulated environments.
