Crossing the last mile: from experimental AI to operational AI in insurance

AI delivers on its promise when it ‘crosses the last mile,’ where innovation finally produces tangible value. But the conditions must be in place to deliver that value. Insurance, often perceived as a conservative sector, is emerging as a prime testing ground for AI, thanks to its large volumes of data, standardised processes and high customer expectations. It has now moved beyond the testing stage to become a real incubator for the operationalisation of AI, constantly seeking value. Use cases are multiplying: automated extraction of complex information, predictive risk analysis, optimisation of claims management processes, automatic quote generation, LLM management assistants, etc. AI is becoming embedded in operations. But human and organisational challenges remain: tech-business alignment, scalability, governance, skills… all essential levers for crossing that last mile. Which uses should be prioritised, which technological choices should be made, and what governance should be put in place for sustainable deployment?