Once AI is confirmed as the right solution, planning ensures algorithm development, data availability, relevance and handling, model management and supporting processes such as quality assurance and configuration management meet the system-level requirements; controls phase transitions with timely error feedback; defines tools, environments and how confidence in them is gained; sets good practice standards for coding, data and test; defines how changes, deviations and waivers are agreed; controls release and integration including how in-situ modification such as online learning is controlled and assured; sets through-life performance monitoring and reporting; manages obsolescence of hardware, software and data; and assesses and reduces environmental impact. Plans are version controlled, agreed in good time, cover tools, models and simulations used for algorithms, data or assurance evidence, and include steps throughout the lifecycle to ensure the ethical principles can be met.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.