SCOR Deploys AI Underwriting Assistant

- SCOR's proprietary "AI Assistant" is a cloud-based, generative AI solution now used daily by over 150 underwriters and claims experts. It processes approximately one million pages per month with around 90% accuracy on key data fields and has already reduced time spent on medical underwriting in their Life & Health division by 30%. - The system is a key component of SCOR's "Forward 2026" strategic plan and was developed in collaboration with Microsoft, which also participated in a hackathon to explore further AI applications in areas like medical impairment detection and automated data capture from P&C documents. - Architecturally, such systems in the insurance industry are moving towards agentic AI and multi-agent ecosystems where specialized AI agents, orchestrated by frameworks like CrewAI, collaborate on complex tasks like risk assessment and decision-making. For backend integration, an API-first approach is often used to connect modern AI tools with legacy insurance platforms, sometimes employing a "strangler pattern" to gradually replace older components without a full system overhaul. - The development of these AI systems often involves LLM orchestration frameworks to manage complex workflows. Popular open-source options include LangChain for building multi-step reasoning chains and ZenML for creating reproducible, production-grade AI pipelines. In the financial sector specifically, open-source models like FinGPT are being developed to cater to the unique needs of the industry. - From a business perspective, the use of AI in underwriting has been shown to improve the combined ratio by up to five points by mitigating fraud before a policy is issued. In claims processing, AI can reduce review times by over 50% and has led to significant reductions in annual underwriting losses for some insurers. - For technical founders in the insurtech space, the current venture capital landscape shows a "flight to quality," with investors favoring more mature, late-stage startups that have a clear path to profitability over early-stage ventures. In 2024, 43% of insurtech VC funding was directed towards B2B SaaS companies. - Recent trends in insurtech funding indicate a rebound in Q1 2025, with a significant portion of capital flowing into startups focused on underwriting and P&C, particularly those leveraging AI for risk modeling and automated pricing. This follows a period of declining investment from a peak in 2021. - Key lessons for first-time fintech founders include the importance of selecting investors who understand the long sales cycles of the financial industry, the critical nature of data quality and infrastructure from the outset, and the need to be patient when dealing with the slower pace of large institutional partners.