The Moving Target: Benchmarking Dynamic AI Systems End-to-End - Gaëtan Brison - Doctolib - dotAI 26

Why watch

Gaëtan Brison presents Doctolib's approach to evaluating complex multi-agent AI systems in healthcare, where isolated component testing is insufficient. The talk introduces synthetic users—AI-powered personas that simulate realistic patient journeys to enable systematic, reproducible evaluation at scale across multiple turns, sessions, and product versions. Designed for AI engineers and ML practitioners building evaluation frameworks for production AI systems in high-stakes domains.