Because AI operates within a wider system that may include other AI, care is taken that it meets its operational requirements where it is actually used, assessing where needed how it bears on operational governance, for instance operational authorities and the rules of engagement. Where performance was proven somewhere other than the final operational system, including transfer to a system with a different build standard, the differences are analysed, understood and mitigated; after substantive modification of the host system continued satisfactory performance is demonstrated; and the approved ODD is clearly defined in the operating context.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.