01
Consequential action
What may bind, change, deny, approve, release, or affect someone?
The system may influence whether applicants advance, are delayed, or are excluded from further review. The requester wants the route evaluated for role clarity, evidence sufficiency, human oversight, logging, and likely EU AI Act high-risk obligations.
02
Requested deliverables
The declared outputs expected from the engagement.
03
Evidence state
Available evidence and declared gaps remain visibly separate.
Evidence available
- System architecture overview
- Model card and intended-purpose statement
- Candidate-screening workflow diagram
- Current human-review procedure
- Sample decision logs
Evidence missing or unresolved
- Validated bias and performance testing record
- Formal Article 9 risk-management record
- Complete data-governance provenance package
- Fundamental-rights impact assessment
04
Required qualifications
Evidence-based qualifications should replace popularity-based reputation.
AI governance or EU AI Act experienceEmployment or workforce-system knowledgeIndependent review capabilityEvidence-mapping and documentation experience
05
Review boundary
The opportunity must not silently expand beyond the declared engagement.
Declared boundary
The engagement requests a bounded governance review. It does not request legal representation, certification, conformity assessment, or a regulator determination.