Capability
Vetted assessment
The talent proves what they can do once. Every company they apply to reads the same verified result, and nobody has to take a resume on faith.
A vetted assessment is a test of role-relevant ability, built to detect cheating, whose result is recorded and reused rather than re-run for every application. On CalHire a talent sits it once: an AI text interview on every role, plus a two-problem coding round with follow-up questions about their own code on technical roles. That produces one vetted score, reused while the result is valid. Companies read the score and the evidence behind it instead of a self-reported claim, and the talent is not re-tested for every job.
Last reviewed
The short version
- A resume is a claim. An assessment result is evidence. The whole design follows from that one difference.
- The vetted assessment is taken once and reused while the result is valid, and can be retaken after a cooldown.
- One vetted score per track: on technical roles the AI interview counts 40% and the coding round 60%, and on other roles the AI interview is the whole score. Every part shows what it rests on.
- A low score never auto-rejects anyone. It changes where a talent sits in a ranked list that a human still reads.
Why the resume stopped working
A resume is a document someone writes about themselves, formatted to please a screening process nobody has explained to them. It reports where someone has been, not what they can do. Those are different questions. Those two things correlate loosely at best, and the correlation is weakest exactly where hiring most needs to be accurate: career changers, self-taught engineers, people who took time out, people whose last company was never going to be a household name.
The screening layer built on top of it has made this worse rather than better. Keyword matching rewards applicants who know which words to use. Generative writing tools have made a polished, targeted, entirely unverifiable resume free to produce at scale, so whatever signal a well-written application once carried has largely evaporated. Recruiters are now reading more applications containing less information.
The alternative is not a better parser. No parser fixes a document nobody has to tell the truth in. It is to stop asking talent to describe their ability and start asking them to demonstrate it, once, under conditions everyone can see.
How the vetted assessment works
One sitting, generated for the talent’s role and declared skills, with the integrity layer on.
- 1
The talent declares their role and skills
Not a work history. The talent states their role, experience level and the skills they claim to have. The skills set the track, technical or non-technical, and the talent can choose the track themselves. Technical roles take the AI interview and a coding round; every other role takes the AI interview, with workplace scenarios among its questions.
- 2
The questions are generated for that sitting
Each interview question is written from the role, the talent’s experience level, their declared skills and the answers so far. On technical roles the two coding problems, one of core logic and one real-world task, are generated for the sitting in a language the talent picks. Nothing is drawn from a fixed public bank, for a practical reason: a static question bank leaks, and a leaked bank measures who found the leak.
- 3
They sit it once, with the integrity layer on
The integrity layer blocks paste and records leaving the page and response timing during the session, and technical roles get follow-up questions about the talent’s own code. There is no proctor watching a video feed and nothing is recorded on video: the camera takes a selfie at the start and two photos during the AI interview, and only people compare them. What is being checked is whether the work is the talent’s own, not whether they looked away from the screen.
- 4
The result becomes one vetted score
The AI interview is rated on four dimensions (communication, depth of knowledge, problem solving and relevant experience), each on a five-level rubric with the answer it rests on quoted. On technical roles the coding round is rated on correctness, approach, code quality, real-world judgement and explaining the code. The interview counts 40% and the coding round 60% on technical roles; on every other role the interview is the whole vetted score. The profile belongs to the talent rather than to one company’s pipeline, so applying to a second job does not mean sitting a second assessment.
- 5
It expires, on purpose
A verified score is valid for a set window. After that it needs refreshing, because a two-year-old result is a claim about the past dressed up as evidence about the present. Talent can retake after a cooldown, which stops score-farming through sheer repetition.
What a company actually receives
One score, with its reasons shown
The vetted score, how the AI interview and, on technical roles, the coding round each scored, and every dimension’s rating beside the answer it rests on. On each job you weigh it against your own interview and the match with the role: 35, 40 and 25 percent by default, adjustable per job.
A date, and an expiry
Every verified score carries when it was earned and when it lapses. A result you cannot date is a result you cannot rely on.
The integrity record
The signals raised during the session, with PII-free evidence attached, so a flagged result can be reviewed by a person rather than quietly discarded.
The same measure, for everyone
Every talent on a track sits the same structure, rated on the same dimensions with the same rubric and weights. The questions differ per sitting; the measure does not. That is what makes a comparison between two people mean anything.
No identity attached
The profile reaches you without a name, photo, age, school or past company. You are reading ability in the literal sense that nothing else is available to read.
A record that survives the decision
Scores, weights and the human decision are written to a hash-chained audit trail. If you are ever asked to account for a hiring decision, the evidence exists.
What makes an assessment worth trusting
The technical term is validity: does the test measure the thing you care about, and does performance on it predict performance on the job. It is a real, measurable property, and most hiring assessments are never checked against it.
Three things move validity in the right direction, and the design follows all three. Content has to be job-relevant, which is why questions are generated from the role and the declared skills rather than pulled from a generic aptitude bank. Conditions have to be consistent, because a test that is easy for one talent and hard for another measures the conditions rather than the talent. And results have to be scored against a fixed rubric rather than an impression.
None of that makes a test perfect. Nothing makes a test perfect. It makes the error understandable, which is a lower bar than perfection and a considerably higher one than most resume screening clears.
What this does not do
An assessment does not tell you whether someone is a good colleague, handles pressure well, or will still want the job in a year. It measures a bounded set of skills at a point in time. Interviews, work samples and references exist because those other things matter, and all three are supported here rather than treated as replaced.
A verified score is not a guarantee against every form of cheating. The integrity layer raises signals for a person to review; it does not deliver verdicts. Any decision that turns on an integrity flag is reviewed by a person, because a false accusation of cheating is a serious thing to get wrong.
A vetted assessment does not by itself make hiring fair. It removes one large source of bias and introduces something new that has to be checked: whether the assessment produces different selection rates across groups. That is what the adverse-impact analysis in the compliance console is for, and it is not optional.
This fits badly where the skill genuinely cannot be assessed remotely in a sitting. Licensed clinical practice, hands-on trades, and roles whose core requirement is a supervised credential need that credential verified first. Assessment is the second step there, not the first.
Questions people actually ask
How long does the vetted assessment take?
Why does a verified score expire?
Can a talent retake the vetted assessment to improve their score?
What stops a talent having someone else take the test for them?
Does the company see the talent’s name during assessment review?
Is the vetted assessment identical for every talent applying to a role?
Related
Where this connects to the rest of the platform.
Assessment integrity without surveillance
Page, paste, timing and identity signals, each with PII-free evidence. No video recording, no verdicts, and environmental flags are excluded from scoring where accommodations apply.
Composite scoring and blind ranking
A job score you set the weights for, defaulting to 35 percent vetted score, 40 percent the job’s own interview and 25 percent match. It orders a list. It never makes a decision.
Anonymous talent screening
Companies evaluate talent with no name, photo, age, school or past company attached, because that information is never sent rather than merely hidden.
Read the reasoning
The evidence and the argument behind what is on this page.
What skills-based hiring actually is (and what it is not)
Skills-based hiring decides on demonstrated ability, not proxies like degrees or job titles. What it means in practice and where it fails.
ReadDoes your assessment predict anything? Validity for non-specialists
An assessment that feels rigorous can predict nothing. What validity and reliability mean, how to check yours, and the four ways hiring tests quietly break.
ReadDropping the degree requirement: what has to replace it
Removing a degree requirement without replacing the signal makes hiring more subjective, not less. What the degree was doing, and what to measure instead.
ReadDoes an ATS really reject your résumé? What actually happens
The "75% of résumés are rejected by robots" claim has no verifiable source. What an ATS actually does, where filtering happens, and what to do.
ReadSee a verified pipeline for one of your roles
Post a role free and review anonymous, skill-ranked talent. No card, no sales call to get started.