In order to ensure a coherent approach, the fundamental rights impact assessment should build upon the risk assessment conducted by providers of high-risk AI systems, where relevant and available.