For every in-scope AI use case the agency must carry out an impact assessment of the use case, begun at the design stage and finalised, with any agreed risk treatments applied, before the solution is deployed. The agency may use the Commonwealth's AI impact assessment tool or an internal process that integrates all of the tool's provisions; an internal process must be consistent and must yield an inherent and residual risk result no lower than the tool would give, and the agency must be able to revise it when the tool is updated. The assessment produces the inherent risk rating that drives the medium-risk and high-risk governance steps and the risk rating recorded in the register.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.