AI candidate evaluation system
An objective, comparable picture of every candidate. IntelliHR’s artificial intelligence measures incoming applications against the requirements of the specific role, requirement by requirement, and gives a comparable picture along six criteria — based on the candidate’s résumé, their pre-screening answers and the position’s requirements.
The evaluation criteria
The AI analysis evaluates candidates along six criteria. Each criterion is classified from a well-defined set of values.
Skills
How well the candidate’s skills cover the role’s skill requirements, weighted by the importance of each requirement.
The weighted coverage of the skill requirements is at least 75%.
The weighted coverage is between 50% and 75%.
The weighted coverage is below 50%.
How do we evaluate?
For each skill requirement of the role, the AI separately judges — from the candidate’s résumé and pre-screening answers — whether the skill is present in the same or only a related field, and whether it is far below, below, at or above the required level. This is an estimate, not a verified fact.
The classification is not given by the AI: the system converts these judgements into coverage with a fixed table, averages them by the weight of each requirement and classifies the result with fixed thresholds. A skill in a related field never counts in full, even at a higher level.
Missing or weakly covered requirements are shown in an explanatory list; they carry no separate deduction. If the role has no weighted skill requirement, the area is not classified (“No requirement”).
Experience
How well the candidate’s professional experience meets the role’s experience requirements, weighted by the importance of each requirement.
At least 75% weighted coverage, and for at least one requirement, experience in the same field above the required level.
The weighted coverage of the experience requirements is at least 75%.
The weighted coverage is between 50% and 75%.
The weighted coverage is below 50%.
How do we evaluate?
For each experience requirement of the role, the AI separately judges — from the candidate’s résumé and pre-screening answers — whether the experience was gained in the same or only a related field, and whether it is far below, below, at or above the required level.
The system calculates the classification with the same rule as for skills. “Exceeds minimum” requires, besides high coverage, at least one weighted requirement that the candidate meets in the same field at a higher level.
Missing or weakly covered requirements are shown in an explanatory list; they carry no separate deduction. If the role has no weighted experience requirement, the area is not classified (“No requirement”).
According to Schmidt and Hunter’s (1998) meta-analysis, years of job experience on their own are a weak predictor of job performance. The evaluation therefore does not look at years on their own, but at whether the experience exists in the required field and at the required level.
Education
How well the candidate’s education and field of study meet the role’s education requirements, weighted by the importance of each requirement.
At least 75% weighted coverage, and for at least one requirement, education in the same field above the required level.
The weighted coverage of the education requirements is at least 75%.
The weighted coverage is between 50% and 75%.
The weighted coverage is below 50%.
How do we evaluate?
For each education requirement of the role, the AI separately judges — from the candidate’s résumé and pre-screening answers — whether the education was obtained in the same or only a related field, and whether it is far below, below, at or above the required level.
The system calculates the classification with the same rule as for experience.
Missing or weakly covered requirements are shown in an explanatory list; they carry no separate deduction. If the role has no weighted education requirement, the area is not classified (“No requirement”).
According to Schmidt and Hunter’s (1998) meta-analysis, years of education on their own are a weak predictor of job performance. Education therefore counts here against the role’s specific requirements, together with the other areas, according to the role’s weights.
Turnover Risk
An estimate of whether the candidate’s employment history so far shows a repeated pattern of leaving early.
A stable employment history, or the résumé gives a reason for short engagements that does not point to repeated leaving.
A mixed picture: both risk signals and stabilising employment history are present.
Repeated, unexplained short engagements without substantial stabilising history.
The résumé does not contain enough employment information for a well-grounded judgement.
How do we evaluate?
The AI assesses the employment history in the résumé as a whole, without computing ratios or averages. A role change at the same employer counts as continuous employment. Fixed-term, project, seasonal and internship work, as well as redundancy or the employer closing down, do not indicate a tendency to leave. Missing dates or reasons never raise the risk; they only lower the strength of the evidence.
Alongside the risk level, the AI also gives the strength of the evidence (weak, medium, strong). Together they can lower the aggregate score by at most 20%; with weak evidence they do not lower it.
Career Progression
The pace and direction of the career path already visible in the résumé. It does not predict future advancement.
Repeated, closely spaced progression along the career path.
Continuous professional progression at a usual pace.
Little vertical movement. This is not a negative signal in itself: stable specialist work can also fall here.
The pace of the career path cannot be judged from the résumé.
How do we evaluate?
Based on the résumé and the experience analysis, the AI judges the path already taken; the target role does not change it, and future capacity is not estimated. Missing dates mean uncertainty, not negative evidence.
The classification is for information only and does not count towards the aggregate score.
Estimated Onboarding Time
An estimate of how long it takes the candidate to work independently in the role’s company-, product- and process-specific environment, assuming the entry prerequisites are met.
Immediate
Can work independently without material onboarding.
Up to 1 week
A short introduction of up to one week is enough.
Over 1 week, under 1 month
A few weeks of onboarding are needed.
1–3 months
Onboarding of more than one month and up to three months.
3–6 months
Three to six months of onboarding.
6+ months
Onboarding of more than six months, justified by the actual difference in the work.
How do we evaluate?
The AI considers directly transferable work experience, the duties that are new to the candidate and the onboarding support offered. The estimate covers role-specific onboarding only, not the time needed to acquire a missing profession, qualification or foundational capability. If there is no well-grounded estimate, the value is “Cannot estimate”.
The AI also gives a confidence level (low, medium, high) and a short explanation with the estimate. It does not count towards the aggregate score.
The relevance score
Alongside the six criteria, every candidate receives an aggregate relevance score (0–100). The score is not given by the AI: the system calculates it with a fixed, traceable formula from the AI’s requirement-by-requirement assessments and the personality questionnaire result, according to the position’s weights.
The personality assessment
If the candidate has completed the personality test, the AI evaluates the results using the Big Five (Five Factor) personality model:
Openness
New ideas, experimentation, attitude toward change.
Conscientiousness
Organisation, planning, discipline.
Extraversion
Drawing energy from social interaction, initiative.
Agreeableness
Seeking harmony, patience, flexibility.
Emotional stability
Stress management, performance under pressure.
For each dimension, the AI measures the result against the personality profile expected for the role and gives a fit value from 1 to 100; the personality area’s value is their average.
The Big Five is a widely used model in personality psychology. Barrick and Mount’s (1991) meta-analysis found that conscientiousness was consistently related to job performance across all occupational groups studied.
Important notes
The AI analysis is a supporting tool; it does not replace the personal interview and the judgement of the recruitment expert.
The evaluation is based on the information found in the candidate’s résumé and pre-screening answers — it does not rest on verified facts.
The personality-test result is based on self-reporting, so it is worth validating the key dimensions during the personal interview.
The system evaluates every candidate with the same methodology and the same criteria, ensuring comparability.
Classifications are to be understood relative to the role — the same candidate may receive different evaluations for different positions.
Referenced research
- Schmidt, F.L. & Hunter, J.E. (1998). The validity and utility of selection methods in personnel psychology: Practical and theoretical implications of 85 years of research findings. Psychological Bulletin, 124(2), 262–274.
- Barrick, M.R. & Mount, M.K. (1991). The Big Five personality dimensions and job performance: A meta-analysis. Personnel Psychology, 44(1), 1–26.
Curious how the AI would evaluate your candidates?
Request a personalised demo — we’ll show you how IntelliHR speeds up your recruitment.