At deployment the team implements the oversight level set at planning, provides a channel for external feedback on possible rights infringements, gives users the information to make informed autonomous decisions, provides the additional process for people with a right not to be subject to solely automated decisions, and trains and supports end-users. It monitors real-world performance and data contamination, concept drift and data drift, keeps the fallback plan ready, documents deployment processes and subcomponents and reviews model documentation. It deploys securely with access limited to required users, documents the system's decisions for traceability, makes available explanations of how far the system shapes decisions and why it was deployed, tells people when they are interacting with AI or receiving AI-assisted outputs, communicates capabilities and limitations, offers a human alternative where rights could be affected, monitors decisions for fairness, engages stakeholders, assesses before deployment that the system's benefit to society outweighs harms, keeps clear accountability so AI decisions trace to people, continues impact assessments, and issues a user guide that is open about strengths and weaknesses and names an ethics contact.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.