FIS has partnered with AI safety company Anthropic to introduce agentic AI capabilities to banking, with the first application focused on combating financial crime.
The collaboration centres on the Financial Crimes AI Agent, which will automate the assembly of evidence across a bank’s core systems at the point a case is opened, assess the activity against established money-laundering typologies, and flag the highest-risk cases for human investigators. BMO and Amalgamated Bank are among the first institutions set to deploy the agent, with wider availability to FIS clients planned for the second half of 2026.
The two firms have taken an embedded approach to development, with Anthropic’s Applied AI team and forward-deployed engineers working directly alongside FIS to co-design the agent and equip the FIS team with the knowledge to build and scale further agents independently over time. Anthropic’s Claude models provide the reasoning layer across the strategy, while FIS supplies the data platform, governance infrastructure, and client relationships. Client data will remain within FIS-controlled infrastructure throughout.
The scale of the problem the agent is designed to address is significant. The UN estimates $2tn in illicit funds passes through the global financial system each year, and US financial institutions alone spend between $35bn and $40bn annually on anti-money-laundering operations. Despite this expenditure, investigators currently dedicate the majority of their time to manually gathering evidence from disconnected systems before meaningful analysis can begin. Emerging US regulation is now pushing institutions to move on from that model, redirecting resources towards the highest-risk activity. The Financial Crimes AI Agent is evaluated on its ability to reduce cost per case, cut low-value manual work, and shorten overall case review time.
FIS sits at the centre of the financial data landscape, acting as the system of record for transactions, payments, deposits, credit, and customer activity across thousands of financial institutions. This positioning gives the Financial Crimes AI Agent native access to the data that investigators require without requiring new integrations or third-party vendor exposure. For institutions running non-FIS core systems, the agent connects via open integration standards. In all cases, the governance, evaluation, and audit layer remains within FIS-controlled infrastructure, ensuring every agent conclusion is traceable and every decision is recorded. FIS describes this architecture as Orchestrated Intelligence.
Financial crimes represents the first application in what FIS has described as a broader roadmap of purpose-built agents on a single governed platform. Further agents covering credit decisioning, deposit retention, customer onboarding, and fraud prevention are planned, each drawing on FIS’s unified data and regulatory infrastructure and each powered by Claude.
FIS CEO and president Stephanie Ferris said, “Every bank in the world wants AI that acts, not just assists. The future is about a trusted provider who manages the data, who governs the agents, and who stands between your customers and the AI making decisions about their money. FIS built the architecture that orchestrates this intelligence. Anthropic is a leading AI provider, Claude is the reasoning engine inside, and the Financial Crimes AI Agent is the first proof of what this architecture can deliver for financial institutions that are ready to become the agent-first bank of the future. It’s a new era in banking.”
Anthropic head of financial services Jonathan Pelosi said, “FIS brings decades of trusted relationships with financial institutions, deep regulatory knowledge, and the transaction data that makes an AI agent useful in practice. That’s why FIS chose Claude, they needed a model that could reason through complex investigations accurately, explain its work, and operate safely inside regulated workflows. We embedded our Applied AI team inside FIS to build the Financial Crimes AI Agent together, so every conclusion the agent reaches links back to its source data, and every decision stays with the investigator.”
Copyright © 2026 RegTech Analyst
Copyright © 2018 RegTech Analyst
