The news: Anthropic is moving beyond being just an AI model provider—positioning itself as a core infrastructure layer for financial work, from back-office compliance to front-office analysis.

Zoom in: Fidelity Information Services is partnering with Anthropic to bring agentic AI into banking, starting with a financial crimes AI agent focused on anti–money laundering (AML). Per a press release:

• The AI agent can automate AML investigations, pulling data across bank systems, analyzing suspicious activity, and prioritizing high-risk cases.
• It aims to reduce investigation time from hours or days to minutes while improving accuracy and auditability.

Early adopters include BMO and Amalgamated Bank, with a broader rollout expected in H2 2026.

More generally, Anthropic is launching AI agents with expanded capabilities:

• Drafting pitch decks for client meetings
• Reviewing financial statements
• Assisting with research and analysis
• Flagging and escalating compliance issues

These tools can integrate directly into many of financial institutions’ (FIs) existing tools such as Excel, PowerPoint, and Outlook and pull data from providers like Moody’s and Dun & Bradstreet.

The risks: Anthropic’s expansion into financial services—launching AI agents that handle analysis, pitch creation, and compliance—comes at the same time regulators are raising urgent concerns about its Claude/Mythos models, creating a clear tension between adoption and risk. 

On one hand, the company is rapidly embedding its technology into core financial workflows and partnering across Wall Street. This signals strong enterprise demand and the potential to reshape knowledge work in finance. On the other hand, regulators and central banks warn these same advanced models could introduce serious risks, from unreliable or opaque decision-making to systemic threats like identifying cyber vulnerabilities at scale or creating shared weaknesses across institutions. 

The result is a “split-screen” dynamic: Anthropic is gaining traction as a foundational AI provider for finance while policymakers are scrambling to assess whether the technology is safe, governable, and resilient enough for widespread use in the global banking system.

Recommendations for banks: FIs must use AI to drive efficiency gains while continuing to build governance, tools, guardrails, and expertise to avoid complete long-term dependency on specific third-party tools. Even as AI automates workflows, accountability remains paramount—meaning human oversight, auditability, and explainability must be embedded from the start.

This content is part of EMARKETER’s subscription Briefings, where we pair daily updates with data and analysis from forecasts and research reports. Our Briefings prepare you to start your day informed, to provide critical insights in an important meeting, and to understand the context of what’s happening in your industry. Non-clients can click here to get a demo of our full platform and coverage.