The How
Performance on real tasks
Scores how candidates actually do the work - video pitches, written plans, live role-plays. Behaviour-level signal, not multiple-choice proxy.
AI Scoring
Explainable AI scoring that earns the hiring decision.
Every score traceable to the evidence that produced it. No black-box rankings, no unexplained rejections - just defensible hiring decisions backed by real work.
How it scores
Three models. One defensible decision.
Each candidate is evaluated across three complementary lenses. No single signal carries the decision - the evidence compounds.
The What
Knowledge that matters
Targeted knowledge checks calibrated per role. Fast to answer, weighted conservatively - used to confirm fundamentals, never to rank on trivia.
The Preference
Style, pace, work-fit
Captures how a candidate prefers to work - collaboration style, pace, ambiguity tolerance. Informs the brief, never the ranking.
Outputs
Made to plug into the hiring workflow.
Auto-ranking
A shortlist you can defend on day one.
Candidates ranked by real evidence, not keyword density. Ties resolved by the weighted skills that matter for the role - never by who submitted first.
Skill group scores
Where each candidate shines - and where they don't.
Scores broken out by skill group, weighted per role. Reviewers see the full picture instead of one opaque number.
AI summaries
A two-minute brief for the interviewer.
Plain-language summary of strengths, watch-outs, and suggested probe questions - pulled directly from what the candidate actually submitted.
Explainability
Every score, traceable to its evidence.
Third-party bias audited
Independently validated by Holistic AI. Fairness results published openly, not hidden behind an NDA. Every model release goes through the same audit.
Consistent every time
The same submission gets the same score - today, next quarter, by any reviewer. Rubrics are stable across the funnel; deterministic where it needs to be.
Defensible to anyone
From candidate appeals to legal review, every score links back to the evidence that produced it. No black-box rejections, no unexplained rankings.
Closed-loop learning
The model gets better because it learns from who actually succeeds.
Most AI scoring stops at the submission. Vervoe keeps watching - four months after hire, the platform checks in with hiring managers and feeds the outcome back into scoring weights. Your funnel improves because the model improves.
Day 0
Candidate submits
Written, video, code, or task-based answers land in the platform. Scoring runs in seconds.
Day 1–30
Recruiter decisions
Shortlist, interview, reject - every recruiter action is logged against the original submission.
Day 30
Hire decision
Who got the offer, who didn’t, and why. The decision trail is preserved alongside the scoring evidence.
~Day 120
Quality-of-hire check
Platform pings the hiring manager four months in: is this person performing? The answer feeds the model.
Ongoing
Model recalibrates
Scoring weights drift toward signals that predicted real-world success - not just assessment completion.
How does Vervoe AI scoring work?
Vervoe AI scoring uses three models: a trained scoring model that evaluates candidates against a structured rubric, a natural language processing model for written and video responses, and a learning model that improves from recruiter feedback and hire outcomes. Every score is tied to specific response evidence - not a black-box output.
Is Vervoe AI scoring explainable?
Yes. Every score on the Vervoe platform includes an explanation - the specific question, the candidate response, and the criteria used to evaluate it. Recruiters can see exactly what drove each score, override it with human judgment, and trace every shortlisting decision back to the evidence.
Does Vervoe AI scoring have bias?
Vervoe AI scoring is evaluated and audited for bias on an ongoing basis. Scoring is based on demonstrated performance on job-relevant tasks - not on demographic information or credential proxies. The platform is Holistic AI Audit certified and designed with bias mitigation requirements in mind.
Can Vervoe AI scoring learn from my team's decisions?
Yes. The Vervoe scoring model improves from human-in-the-loop feedback. When your recruiters override AI scores or provide calibration signals, the model learns from those decisions and becomes more accurate for your specific roles and standards over time.
See what explainable AI scoring actually looks like.
Role-specific. Evidence-backed. Open to audit. Stop guessing which candidate will perform - know.