OPS.11:5.3 - AI-assisted operation: generation, acceptance and deployment

An AI-assisted software operation can generate candidate changes faster while human acceptance, test environments or deployment permissions remain unchanged. Candidate changes, tool calls, episodes, evidence and accepted releases have different identities and resource demands.

The practitioner follows only the relevant couplings: which configuration a test supports, who can accept the change, which provider/tool access is available, and what deployment conditions apply. A useful result might reduce admissions to the acceptance capacity, reserve a compatible test environment and return an unresolved release decision. It does not infer agent capability or software assurance from a higher generation count.