01 Executive summary
You have a clear view of where AI sits across the HR lifecycle. That puts you ahead of most teams at this stage.
Five of the eight areas use AI today; three are high risk under the EU AI Act. Emotion analysis in video interviews is prohibited and needs to come off this month.
The rest is procedural: a named reviewer for screening, a worker notice, and vendor documentation on file. None of these require a tool change. Starting now gives you time to sequence them well across the year.
02 Readiness across six capabilities
- AccountabilityAct now
A named person can review and overrule each AI-influenced decision.
- TransparencyPrioritise next
Candidates and employees are told when AI is shaping decisions about them.
- ExplainabilityPrioritise next
You can explain the AI's role in a decision if asked.
- Record-keepingPrioritise next
Logs and oversight records are kept and exportable on request.
- Vendor due diligenceAct now
Vendor roles, documentation, and ongoing assurance are in place.
- AI literacyOn track
People using the tools have role-appropriate training.
- Candidate Screening. Automated decision-making on people. High-risk under Annex III(4) and within scope of Art. 22 / Art. 86.
- Interview Assessment. Inferring emotion from workers or candidates is banned in the workplace under Art. 5(1)(f). Stop the use now.
- Performance Evaluation. AI narrows the field and rejected people are not meaningfully reviewed. That is a material-influence decision.