Allianz partners with Anthropic to scale responsible AI in insurance

Allianz and Anthropic have formalized a global partnership aimed at scaling the use of responsible, auditable AI across Allianz Group, according to a press release dated January 9, 2026. The collaboration is positioned around high-impact insurance use cases, where AI outputs can directly influence operational decisions, claims handling, and customer outcomes. The primary audience includes Allianz employees, developers, operations teams, and ultimately customers interacting with AI-supported insurance workflows.

“Insurance is an industry where the stakes of using AI are particularly high: the
decisions can affect millions of people. Allianz and Anthropic both take that very
seriously, and we look forward to working together to make insurance better for those
who depend on it,” said Dario Amodei, CEO and Co-Founder of Anthropic.

At the core of the partnership is the integration of Anthropic’s Claude models into Allianz’s internal AI platform, making them broadly available to staff across business units. Allianz developers are already using Claude Code for software development tasks, while upcoming integrations using Model Context Protocols are designed to allow secure connections between Claude and Allianz’s internal systems and data sources. This setup is intended to support enterprise use at scale while maintaining strict controls over data access and model behavior.

Beyond general access, the two companies are jointly developing custom agentic AI systems tailored to Allianz’s operational needs. These agents are designed to handle multi-step workflows such as claims intake, document classification, and structured case preparation, with early focus areas including motor and health insurance. The agents are explicitly built with human-in-the-loop mechanisms, ensuring that sensitive, ambiguous, or high-risk cases are reviewed by Allianz staff rather than handled autonomously.

A third pillar of the partnership focuses on governance and compliance. Allianz and Anthropic are co-developing traceability and observability features that record how AI systems arrive at decisions, including inputs, intermediate reasoning artifacts, and outputs. These logs are intended to support regulatory oversight, internal audits, and post-hoc analysis, reflecting the requirements of operating AI systems in heavily regulated insurance environments.

For Allianz, the partnership extends its existing AI initiatives, which already include multilingual customer assistance and reduced processing times for certain claims and invoice workflows. The collaboration emphasizes controlled deployment, transparency, and accountability rather than full automation. For Anthropic, the deal represents a deeper move into regulated enterprise environments, positioning Claude as a model suitable for organizations where safety, explainability, and compliance are non-negotiable constraints rather than optional features.

Alongside this partnership, Anthropic announced that Roman Howe has been appointed Head of DACH and CEE. Roman joins from his role as Chief Revenue Officer at FactFinder and will be based in Anthropic's recently announced Munich office.