“How do you govern the AI used in this service?”
A buyer can ask that question in one sentence. The answer often sits across a spreadsheet, a supplier review, policy documents, tickets and email. The practical problem is not producing more files. It is showing how the current system record, accountable owner, assessment, evidence and unresolved work fit together.
Recreated interface using synthetic data. Illustrative indicators are not legal compliance scores.
What the six screens show
- Capture the system and owner. Record purpose, context and the person accountable for keeping the entry current.
- Review the gaps. Separate what is known from what still needs investigation or a decision.
- Connect evidence. Link supporting material to the record it supports, with its source and history.
- Assign actions. Turn an unresolved gap into owned work with a clear next step.
- Report the current position. Bring the maintained record together for an internal review, board update or buyer question.
- Support contributors. Use Academy learning to help people understand and maintain the process.
This is the working pattern behind our guidance on building an AI inventory and turning a risk register into owned action. Software can organise the record and workflow; accountable people still make the decisions.
A buyer-ready answer should be concise without hiding uncertainty. State what the system does, who owns the decision, which evidence supports the current position and what remains open. If a control has not been tested, or a supplier document is out of date, record that as an action rather than rounding it into a reassuring score. That makes the answer easier to maintain and more useful when a board member, buyer or reviewer asks what changed.
Fictional example
System: an assistant used to draft internal support FAQs. Owner: Alex in Operations. Evidence: a supplier review, approved-use note and staff review procedure. Unresolved action: confirm the next supplier-model review date and assign it to the service owner.
When a buyer asks how the service is governed, the response can point to a named owner, the relevant evidence and the open action. It does not turn a fictional percentage into a compliance score or claim a measured customer outcome.
Request a walkthrough
Explore the AI Act Ready software workspace if your team needs one maintained place for inventory, risks, controls, evidence and actions. If you need help setting up ownership, assessments or the operating process, see our AI governance consultancy.
AI Act Ready provides operational governance support, not legal advice, certification or a guarantee of compliance.