How is an AI candidate score actually calculated, and what does a score like 87 out of 100 really mean?
A candidate score is calculated by comparing the information extracted from a CV with the requirements of a specific job. Experience, skills, seniority, responsibilities and other role-specific criteria are evaluated together to measure how well the profile fits the position. The score summarizes that fit, while the reasoning behind it shows the recruiter why the candidate received that result and what should be explored further during the interview.
Why is an unexplained score a problem?
A number on its own pushes people toward two equally risky behaviors: either blindly trusting it and ruling out a candidate without checking why they scored low, or ignoring it completely because it isn't trusted, which makes the tool pointless. Both reactions come from the same root cause: the score arrived without its reasoning attached. An explained score avoids both extremes. Recruiters can check the logic, challenge it if needed, and make an informed decision rather than blindly following or ignoring a number. This is the difference between a scoring tool that saves time and one that just adds a layer of unearned authority to the process.
What criteria does the score actually rely on?
Candidate scoring usually draws on several combined dimensions: how well the candidate's profile matches the specific requirements of the job description, the strengths identified in their background, and the points worth digging into during the interview rather than ignoring. It's not a simple average of keywords found in the resume, but an evaluation that accounts for the context of the target role, which is why the same profile can score very differently depending on the job. A candidate who scores 90 for a senior technical role might score 60 for a junior generalist position, and that difference is the whole point: the score reflects fit, not a generic quality rating.
| Scoring dimension | What it means |
|---|---|
| Job fit | How well the candidate matches the specific requirements of the role |
| Strengths | The main elements in the candidate's background that support their fit |
| Points to explore | Gaps, unclear elements or topics to verify during the interview |
| Role context | The score depends on the specific job, not on a generic candidate rating |
How is job fit actually measured?
Job fit isn't calculated by counting how many words from the job posting also appear on the resume. That approach rewards resumes written to game a keyword filter, not resumes that reflect real experience. Instead, the requirements in the job description are broken down into what they actually mean: the seniority level implied, the specific responsibilities listed, the tools or methods mentioned. The candidate's background is then read against that breakdown, not against a flat keyword list. A resume that describes leading a team without ever writing the word "manager" can still score well on a leadership requirement, because the evaluation is reading for substance, not for exact phrasing.
What strengths and interview points come out of the score?
The score isn't just a single figure sitting on its own. Alongside it comes a short list of the candidate's identified strengths and, just as importantly, the points worth a closer look during the interview. These aren't weaknesses in a negative sense: they're the gaps or ambiguities the resume didn't fully answer, the kind of thing a recruiter would naturally want to ask about anyway. Having them surfaced ahead of the interview means less time spent re-reading the resume line by line and more time actually preparing the right questions for that specific candidate.
Does the score change from one recruiter to another?
No, and that's actually one of the main benefits of this kind of tool. Unlike a human read, which varies depending on the person, their fatigue, or their mood that day, candidate scoring applies the same criteria consistently to every application reviewed for a given role. This consistency doesn't replace the recruiter's final judgment, but it guarantees a fair starting point before the recruiter steps in with their own perspective. Two recruiters looking at the same profile on different days will see the same score, even if their personal read of the resume differs.
Can candidate scoring be biased?
Like any system, it depends on the criteria defined upfront. A score built on vague or poorly defined requirements will reflect that vagueness. This is exactly why the reasoning behind each score stays visible instead of hidden behind an opaque number: a recruiter who sees an unexpected result can trace it back to the actual criteria and judge for themselves whether the logic holds up for that candidate. Bias in scoring is rarely about the tool itself, it's almost always about how the job requirements were framed in the first place, and documented reasoning is what makes that visible and fixable.
Where does candidate scoring fit in the hiring process?
Scoring doesn't happen in isolation. It comes after resumes have already been collected and sorted into a ranked shortlist, the kind of output that ATS integration produces when it connects your applicant tracking system to the rest of your recruiting stack. Scoring is what turns that shortlist from a simple ranking into something recruiters can actually reason about: not just who's on top, but why.
Does a high score mean the resume review stops there?
Not quite. The score is built on top of a structured read of the resume, the same structured breakdown that CV analysis produces when it extracts a candidate's experience, skills, and background into something consistent and comparable. Candidate scoring is the next layer on top of that groundwork: it takes the structured data and evaluates it against a specific role. A high score is a strong signal to prioritize that resume for a closer human read, not a substitute for one.
How does MONARK PEPIT document its scoring?
At Monark Pepit, every score comes with its reasoning built in: job fit, identified strengths, and points worth a closer look in the interview. Nothing is presented as raw, final truth. Everything stays available in detail, which lets recruiters challenge a surprising result instead of accepting it without understanding where it came from. That documentation is what turns the score from a black box into a working tool the team actually trusts over time.
Common mistakes teams make with candidate scoring
A few habits undermine the value of scoring even when the underlying tool works well. Treating the number as a final verdict instead of a starting point is the most common one: a low score should prompt a quick check of the reasoning, not an automatic rejection. Comparing scores across different job descriptions is another: an 80 for a senior role and an 80 for a junior role aren't measuring the same thing, so ranking candidates from different postings against each other on raw score alone doesn't hold up. Finally, leaving the underlying job description too vague weakens every score that gets calculated from it, since the model can only evaluate fit against what was actually specified.
Keeping the score useful as hiring volume grows
As the number of applications per role increases, the temptation to skip straight to the top-scored candidates gets stronger, and that's precisely when the documented reasoning behind each score matters most. It's what lets a recruiter move quickly through a large volume of applications without losing the ability to sanity-check a result that looks off. Combined with the rest of a recruiting stack, candidate scoring holds up well as one more layer in an HR tools setup built for volume, not as a standalone gatekeeper making decisions on its own.
FAQ
Does a high score guarantee a good hire?
No, the score remains a screening aid, not a guarantee of success. The final decision, including the interview and human evaluation, remains essential.
Can candidate scoring be biased?
Like any system, it depends on the criteria defined upfront. That's exactly why the reasoning is documented and available, rather than hidden behind an opaque number.
Can scoring criteria be adjusted for a specific role?
Yes, the score is calculated based on the provided job description, which allows the evaluation to be adapted to the real priorities of each hiring process.
Is candidate scoring the same as an ATS ranking?
Not exactly. An ATS ranking usually sorts applications by basic filters or keyword matches. Candidate scoring goes further, evaluating fit against the actual requirements of the role and explaining the reasoning behind each result.
Does the score account for soft skills?
Where the resume gives signals of them, such as described leadership or communication experience, yes. But soft skills are also one of the areas the score flags as worth confirming directly in the interview rather than assuming from the resume alone.
Discover MONARK PEPIT, AI-powered CV screening and interview preparation for recruiters.
