What the AI never does

For a recorded answer or AI interview, ImpactHire scores what the candidate said. It does not use how the candidate speaks to calculate the score.

The candidate's face, voice, accent, manner, or presentation style does not contribute to the score.

This distinction matters because characteristics of speech can reflect a person's health, first language, or background. ImpactHire therefore scores the words in the answer and leaves the rest outside the scoring calculation.

For an AI interview, the recording and transcript remain available as part of the record, but the score itself reads the words alone.

ImpactHire produces scores and the evidence behind them. It does not produce a recommendation about what your company should do with a candidate.

A score can be configured to move a candidate forward to another workflow stage when it passes a threshold.

A score cannot automatically reject a candidate. If a candidate's score does not reach the threshold, they remain in their current stage with the evaluation available for your company to review.

ImpactHire also does not decide whether to make an offer. Hiring decisions remain with the company.

Integrity events follow the same principle. An integrity flag reports something the product observed. A person decides what that observation means.

AI-generated scores are reviewable and overridable.

If a person disagrees with the level assigned to a criterion, they can select a different level and give a reason. ImpactHire then recalculates the answer score and total using the new level.

The original machine score is not erased. ImpactHire keeps the original level and reason alongside the person's override, including who made the change and when.

This means the final figure can reflect human judgment without removing the record of what the AI originally produced.

Still stuck? Contact support.