How HireSalt scores a resume, and why you can read the reasoning behind it

HireSalt gives every resume a score from 0 to 100. This article explains how AI resume scoring works in HireSalt, using one candidate from start to finish. The score is a weighted average of category scores. Each category breaks down into sub-criteria, and each sub-criterion comes with a short note from the AI. You can read those notes, change the weights, and see what moves.

Candidate A is made up for this article. The numbers are illustrative and are not output from a real screening run.

What happens after you upload

HireSalt pulls the text out of the PDF or DOCX file. The AI reads that text next to the job description and your criteria weights, then scores each category and each sub-criterion.

Two things improve results. A longer, more detailed job description gives the AI more to score against. Two limits on reading files: tables inside Word files are not read, so skills listed only in a table are missed, and scanned PDFs are not read.

Seven categories, five of them weighted

HireSalt scores seven categories. Five carry weight by default and count toward the final score: Skills Match, Experience, Education, Semantic Fit, and Stability & Trajectory. Practicalities and Engagement Profile are scored and shown as informational. They carry zero weight, so they do not change the total.

By default every sub-criterion is switched on. The five weighted categories hold 25 of them, and the two informational categories hold another 8, which makes 33 in all. Skills Match has five, including Required Skills Coverage. Experience has seven.

Candidate A’s score (illustrative)

Assume a mid-level role on the default split. Candidate A’s category scores look like this:

Contribution is weight times score. Practicalities and Engagement Profile do not move the total, but their notes still deserve a read. They can raise questions the number does not show.

This example uses the standard categories only. If your job also has criteria the AI drafted from the job description, or custom criteria you added, each set is averaged on its own and the final score blends the sets by a split. A job with AI-drafted criteria and no saved split counts the standard categories for 70% and the AI-drafted set for 30%. In the criteria editor, the AI-drafted and custom sets get 10% to 25% each, depending on how many criteria they hold, and the standard categories keep at least 50%.

You change the split under Scoring Criteria in the Scoring Setup tab. Set it before you upload resumes. Once HireSalt scores a resume for that job, the weights lock. To try another split, clone the job or create a new one.

Reading the sub-criteria

Open Skills Match and you see five sub-criteria. Each one has a score and a note of one or two sentences. Candidate A’s note on Required Skills Coverage might say that four of the five must-have skills appear in the resume and that Kubernetes is never mentioned.

The note is the AI’s summary of what it found and what it missed. It is not verified evidence. Check it against the resume before you act on it. A gap like the missing Kubernetes line also makes a good interview question.

The radar chart

The radar chart draws one axis per category. A larger shaded area means a stronger overall profile, and an uneven shape shows where the gaps are. The dotted line marks the threshold, which defaults to 50.

For Candidate A, Education and Stability & Trajectory, both at 80, reach furthest out. Experience, at 60, is the weakest of the five weighted categories. In the Side-by-Side tab you can overlay up to three candidates on one chart.

Heatmap and quadrant views

The heatmap puts candidates in rows and criteria in columns. Each cell is colour-coded from strong to weak, so you can see which criteria struggle across the pool.

The quadrant view plots each candidate on two axes that you pick from the dropdowns. The midpoint on each axis is 50. In the Skill × Experience view, the top-right corner is labelled Ideal Candidates. Candidate A, at 70 on Skills Match and 60 on Experience, lands there. Views that plot Practicalities or Engagement Profile stay hidden under the default split, since those categories carry no weight.

Shortlist check: test the weights

The Shortlist check tab has a weights check. It moves each weight up to 10 points above or below your job’s baseline and names anyone who would come above your interview line with that small change. No new AI scoring run starts. Nothing is re-scored and no stored score changes.

If the check names someone, your list depends on the exact weights you picked, so read that candidate before you shortlist. If nobody moves, the order did not depend on a small change in emphasis; that does not show the ranking is correct.

Adjust weights opens one slider per category. The ranking and your line update as you drag, and a Who moved list shows each candidate whose rank changed and anyone who crosses the line. A ranking can look precise while several candidates are nearly tied, so read the people who move before you trust the order.

For more detail, see how AI scoring works, criteria and sub-criteria explained, and the weights check in the help center.

Try it on a real role

Pick a job you are hiring for now. Set the weights, then upload a small batch of resumes. Open the top three and read the note behind each one’s weakest category. Open the Shortlist check tab and read the weights check. If nobody moves, the order did not depend on the exact weights. If someone does, the candidates who move are the ones to read first. Upload one open role and see how the breakdown reads for your own shortlist. New accounts start with free credits; packs and plans are on the pricing page.