Human-First AI

AI that assists, not decides

AI supports consistent simulation delivery and can identify relevant evidence and suggest a score against approved behavioural criteria. A qualified reviewer determines the final rating.

Evidence you can inspect

Suggested scores are presented with the supporting transcript evidence and rationale, so reviewers can easily examine how each suggestion relates to the assessment criteria.

Clear boundaries

Transcript-based scoring focuses on what a participant says, proposes, decides, recognises, challenges and commits to. It does not assess appearance, body language, accent, pitch or tone of voice.

Consistent application of criteria

Each skill is evaluated separately against the same approved behavioural indicators and scoring anchors. This keeps the focus on the relevant evidence for that skill.

Meaningful human oversight

Reviewers can confirm, amend or reject suggested scores; remove or add evidence; and review recordings where enabled. A suggestion cannot become a final rating without human review.