AI moving into underwriting
JPMorgan’s Jamie Dimon flagged AI as reshaping underwriting and risk functions with hundreds of use cases — he cited about 600 applications and 50 critical ones that could compress workflows dramatically. Captive insurers are already using AI to screen submissions and flag outliers, and KPMG warns fast claims automation still needs transparent governance to maintain trust. ( )
JPMorgan says it now spends roughly $2 billion a year on AI as part of an $18 billion technology budget. (finance.yahoo.com) The bank’s internal LLM Suite was rolled out to more than 200,000 employees within months and connects to external models (OpenAI, Anthropic) inside a secure corporate portal. (jpmorganchase.com) JPMorgan executives have shifted from pilots to production, reporting several hundred active AI deployments and public reporting that the firm is tracking ROI at the initiative level while targeting expansion toward 1,000+ use cases. (tearsheet.co) Captive insurers are already adopting AI for underwriting and operations, with Marsh’s ReadyCell marketed as an AI-enabled product that can open a cell captive in minutes and panels at captive conferences highlighting AI for pricing, underwriting and governance. (captiveinsurancetimes.com) Commercial insurers are applying AI at submission intake to triage documents, score appetite alignment, flag compliance or outlier submissions, and detect pre-bind misrepresentation and fraud before policies are issued. (iireporter.com) KPMG’s industry report warns that rapid claims automation raises trust, accuracy and security concerns and recommends formal governance, human-in-the-loop controls and an assessment framework to guide insurer rollouts. (assets.kpmg.com)
