The intended decision is which sites and topics deserve limited audit or engagement resources.
Proposed mechanism: Target inquiry according to relevant risks instead of one questionnaire for all.
Limit: Targeting requires validated criteria and may still miss unmodelled harms.