The headline change in this release is structural: workloads now model as a three-level hierarchy of systems, user journeys, and services, rather than a flat list. On top of that sits generative-AI failure-mode analysis that reasons about how a workload can break and checks it against Well-Architected best practices, so you get suggested weak points instead of a blank assessment to fill in yourself. The other substantive piece is dependency discovery. Assessments now crawl the actual dependencies a workload leans on — AWS services plus internal and third-party endpoints — which is the part platform teams usually reconstruct by hand. Resilience policies are modular, and org-wide reporting rolls up through AWS Organizations. It is available in every Region where the service already runs, so there is no rollout to wait on.