SCOR Deploys GenAI for Underwriting & Claims

SCOR's "AI Assistant" moves beyond simple automation by creating a "digital twin" of applicant data extracted from unstructured medical records and reports. This structured data is then used to augment human underwriters, who remain in control to supervise and validate the AI's output. The system, built internally by and for underwriters, is designed to analyze thousands of pages in seconds and is already used daily by over 150 of SCOR's underwriting and claims experts. The collaboration with Verisk on the EHR Triage Engine aims to accelerate underwriting by digitally analyzing consumer-authorized electronic health records, reducing a process that can take weeks down to minutes. This tool, which leverages advanced predictive analytics, can assess 95,000 impairments and qualify up to 85% of applicants with minimal human review. The engine integrates via a standalone API or directly into SCOR's Velogica automated underwriting system. In Vietnam, SCOR's partnership with Insmart JSC deploys a predictive engine developed by its "BeeTech" solutions platform. BeeTech focuses on risk quantification using machine learning and is part of SCOR's broader "Digital Solutions" platform, which also includes the Velogica underwriting suite and the VClaims digital claims solution. This modular, SaaS-based platform is designed for global scalability, allowing for tailored integrations with client systems. Architecturally, such systems represent a shift from monolithic AI to multi-agent systems, where specialized AI agents handle distinct tasks like intake, document analysis, and fraud detection. This agentic approach mirrors the functional structure of human claims departments and allows for greater scalability and adaptability. The orchestration layer becomes critical, managing the sequence of tasks, ensuring data flows between agents, and incorporating human-in-the-loop validation for complex decisions. For platform engineers, the rise of agentic AI reframes API management from a simple connectivity layer to a strategic control plane for the entire AI lifecycle. This "AI supply chain" relies on robust API gateways to manage access to data, orchestrate model interactions, and enforce governance policies. The architecture must support both model-centric APIs (for specific tasks like data extraction) and agent-interaction APIs that manage conversational context and orchestrate multiple models. Venture funding in insurtech has seen a market reset, with a significant drop from its 2021 peak. However, investor focus has pivoted sharply toward B2B SaaS solutions, which captured 43% of insurtech VC funding in 2024, the highest share ever. While overall deal volume is down, early-stage startups are seeing larger median deal sizes, indicating selective but strong backing for promising foundational technologies. For technical founders, this environment underscores the importance of a strong sales capability alongside technical expertise. While venture funding is an option, building organically with a focus on a specific market need is a viable path. When seeking investment, having Directors & Officers (D&O) and Errors & Omissions (E&O) insurance in place is often a prerequisite from investors to mitigate leadership and product liability risks.