You set the screening rules on every role
Job match and resume flags are advisory call prep. Pre-screening knockouts and criteria you configure can filter applicants before assessment when you enable auto-reject on the role.
What we do
- Job match scores and resume flags are advisory. They surface call prep and do not auto-reject candidates on their own.
- Knockout questions and pre-screening criteria you set on a role can auto-filter applicants when auto-reject is enabled. Rules you control per job.
- Proctor assessments and attach integrity context to verified scorecards.
- Show four separate signals on every candidate: Session Confidence, Skill, Job Match, and Identity.
- Encrypt candidate PII. See /data-security.
What we do not do
- Auto-reject from job match score or resume flags alone.
- Auto-reject anyone from an identity check, or let it change Skill or Job Match scores.
- Screening behavior follows what you configure on each role. No platform-wide letter grades (A/B/C/D) or hidden AI cutoffs.
- Claim to be a consumer reporting agency or FCRA background check.
- Market resume flags as definitive fraud or "lie detection" claims.
Do you auto-reject candidates below a match score?
Not on match score alone. Job match is advisory call prep. If you enable pre-screening auto-reject on a role, applicants who fail knockout questions or criteria you configured can be filtered before assessment.
Who decides who gets rejected?
Your team sets pre-screening rules per job (knockouts, must-haves, optional auto-reject). Match scores and resume flags inform your calls. They stay advisory, not a hidden platform cutoff.
Who judges assessment responses and scores?
For open answers, an LLM scores against the role brief, the question, and a structured rubric for that format. Rapid fire uses answer keys. Hands-on and Conversation have their own rubrics. Proctoring attaches integrity context. A recruiter still decides who advances. See /scoring-methodology and /transparency.
What are the four signals on a candidate?
Session Confidence shows how cleanly the assessment ran under proctoring. Skill is the score for the work the candidate did. Job Match compares the resume with the role. Identity shows whether public professional sources back up who the candidate says they are, when your company runs identity checks. Each one is scored on its own. How each is built: /scoring-methodology.
How is the Identity signal used?
Identity is a separate, advisory signal. It never feeds Skill or Job Match, never rejects anyone on its own, and is visible only to the hiring team for that role. Use it as a prompt to ask the candidate a question. Details in our Privacy Policy: /privacy
Why should we trust AI-generated scores?
Because Skill scores come only from demonstrated work under disclosed proctoring, and humans can open the transcript and disagree. We do not claim cheat-proof or bias-free. We claim inspectable decision support. Match and resume flags stay advisory. Details: /transparency and /trust-and-scoring.
How do you reduce bias in AI hiring?
Skill scores come from proctored performance alone, with no weight on pedigree, school, or name. Identity checks sit in their own signal and never touch Skill or Job Match. Every decision stays with a recruiter: job match and resume flags are advisory and never reject anyone on their own. When you use our assistant to turn a plain-language request into pre-screening rules, it rewrites phrasing that could be discriminatory into job-related criteria, and automated rejection messages avoid protected characteristics.
Are you bias-free or EEOC certified?
No hiring system is bias-free, and we do not claim a certification. TalentProof is decision support: verified skill measurement plus a human in the loop, so a model never makes an adverse call by itself. See /scoring-methodology for how each format is scored.
Bias and fairness
The real bias risk with hiring AI is a model quietly filtering people out on proxies for age, gender, race, school, or name. We reduce that risk two ways. Skill scores come from proctored performance alone, with no weight on pedigree, school, or name, and the Identity signal is kept separate from Skill and Job Match. And we keep a recruiter in every decision, so match scores and resume flags stay advisory and never reject anyone alone.
When you use our assistant to draft pre-screening rules from a plain-language request, it rewrites phrasing that could be discriminatory into job-related criteria, and automated rejection messages never reference protected characteristics. This is decision support, not automated selection. No system is bias-free, so we do not claim to be.
Why this matters
Regulators and hiring teams scrutinize automated hiring scores. TalentProof is built as decision support for recruiters, especially on early funnel filtering and verified skill proof.
Every candidate gets four signals, each shown on its own: Session Confidence, Skill, Job Match, and Identity. None of them rejects a candidate by itself.
For how each assessment format is scored, see /scoring-methodology.
Related
Questions on compliance?
Contact us for security questionnaire or partnership review.