Explainability means the main factors that shaped a decision can be stated in terms humans understand. It matters most where decisions affect people, who distrust a decision they cannot follow, especially an adverse one such as a refused loan, and it also validates the system where decisions do not touch humans, for instance by exposing that an image classifier relied on spurious correlations in training data. The organisation accounts for the fact that deep neural networks are hard to explain, that rule-based and decision tree methods are more explainable, and that even those become less understandable as they grow.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.