Every mature regulated sector has developed a specific role for the person who is personally accountable for governance document quality. Not the team. Not the process. A named individual with authority, a defined standard, personal liability, and an external body that can hold them to it.
The actuarial profession has the appointed actuary. Under ASOP 56, the actuary who signs off on a model must document their work in a form that allows another qualified actuary to assess its reasonableness. They must disclose material limitations, reliance on others, and known weaknesses. Personal professional liability attaches to what they certify. An external board enforces the standard. Management cannot override it.
The pharmaceuticals industry has the Authorised Person. Under PIC/S GMP requirements, the AP must personally certify every batch before it can be released for sale. The role cannot be delegated to anyone who is not also a qualified Authorised Person. The AP must be permanently available to the manufacturing authorisation holder. No batch ships without that signature.
Aviation certification has the designated engineering representative. Nuclear safety has the safety case author, named and accountable under the safety case principles.
Each role shares the same structural logic. The individual carries personal professional liability, not just organisational liability. Their accountability is defined by an external standard that management cannot override. And their sign-off is a prerequisite for deployment, not a courtesy.
AI governance has no equivalent. The people who write the governance documents (the system prompts, the deployment policies, the evaluation criteria) have no formal accountability, no external standard, and no recognised professional role. If the documents are wrong, no named individual is responsible for their structural quality. That is not a technology problem. It is an institutional design problem that every other regulated sector solved decades ago.
PromptQ does not create the role. It provides the instrument the role would use. But the accountability question remains open.
Who is the Authorised Person for your AI governance documents?