Explainability is the ability of an AI system to express, in a way people can understand, the important factors behind its results, and some outputs are harder to explain than others because of complexity (deep neural networks more than decision trees; ISO/IEC 22989, 5.15.6). For each identified interested party the organization should analyse the foreseeable benefits and harms tied to explainability, for example: whether the information needed to make informed decisions from the outputs is available, where missing information matters for the use case; how explainability is handled at life cycle stages such as design and development, verification and validation, and deployment (whether it helps improve the deployed system or decide on adoption); its effect on interested parties' trust; and the effects of failed, wrong or incomplete explanations.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.