Start with the source
A supported document scan can propose a finding and point to a passage. That proposal is not verified evidence until a person reviews and confirms it.
Work requirement by requirement
The Frameworks register shows requirement state, supporting evidence and next action. A deterministic, versioned engine makes historical snapshots traceable to the rules used.
Keep the decision human
A reviewer links evidence, records justification and addresses findings. Readiness is an internal work signal, not a legal verdict or certification.
Example: a finding that changes the work queue
A scanner points to a policy passage that may leave human oversight unclear under a selected EU AI Act requirement. The reviewer checks the passage and system context, links an existing procedure or asks the owner for missing evidence, then records the decision. Fencia recalculates the requirement state from these inputs. The number may change because the evidence changed, not because an AI model declared the organization compliant. A later snapshot retains the registry version used for that calculation.
A repeatable review sequence
- Confirm the AI system, role and requirement before treating a finding as applicable.
- Open the cited passage and inspect the full underlying document where needed.
- Confirm or reject proposed evidence, attach the reviewer and explain the decision.
- Follow the next action and compare later snapshots using their registry versions.
Common questions
- Is readiness a compliance score?
- No. It is a bounded internal signal calculated from requirements, evidence and decisions. Applicability and sufficiency still require qualified review.
- Will a model finding count as verified evidence?
- No. A person must confirm the evidence before the governance engine or Trust Center treats it as such.
Bring the question into a review
Explore a fictional Fencia walkthrough or tell us which system and evidence your team needs to review.