AXA

AXA faced challenges managing vast amounts of unstructured data (image, text, voice) across insurance operations. Underwriting analysis took months or weeks to process historical data, and call center agents required an average of five minutes to locate policy information to answer customer queries.

AXA deployed AI across multiple insurance workflows: AI-powered underwriting risk assessment to accelerate data analysis, real-time dynamic pricing, call center AI assistants that surface policy references within seconds of a customer question, call summarization tools, and parametric insurance automation that triggers automatic claim payments when predefined conditions are met.

Underwriting analysis time dropped from months/weeks to hours/minutes. Call center query response time fell from ~5 minutes to ~5 seconds per policy lookup, significantly improving customer satisfaction and retention. Industry benchmarks cited by AXA's CIO show claims processing automation cuts admin costs by up to 30%, risk assessment accuracy increases by up to 50%, and policy monitoring reduces review times by 70%.

• Focus AI investment on a small number of high-impact use cases aligned to business strategy, rather than adopting every available tool and accumulating technical debt.
• Explainability and regulatory compliance (e.g., EU AI Act) must be built into AI solutions from the start, especially for pricing and risk assessment.
• Human agents remain essential for empathetic edge cases; AI should augment service channels rather than replace human judgment entirely.