The procedure required by the EU AI Act for providers to demonstrate whether a high-risk AI system fulfills all the mandatory legal requirements outlined in the Act. This must be carried out before the system is placed on the market or put into service, and may involve either internal control or a Notified Body.