Step 3 · Fit ranking
Rank every candidate on evidence, not gut feel
AIcruiters screens and ranks each sourced candidate against the complete vacancy. Every match score comes with its evidence: the experience that fits, the requirements that are missing and the risk factors worth checking. Recruiters see why someone ranks high, can challenge the outcome, and low-confidence cases go to a person instead of being processed automatically.
What is wrong with screening profiles by hand?
A recruiter reading a long list of profiles makes dozens of small judgements an hour. The first profiles get careful attention, the last ones a glance. A familiar employer name pulls someone up, an unusual title pushes someone down. None of that is written down, so a colleague or a client cannot see why candidate twelve made the shortlist and candidate thirteen did not.
Keyword screening does not solve this. It rewards profiles that repeat the right words and misses people who describe the same experience differently. What a recruiter needs is the same careful reading applied to every profile, with the reasoning visible afterwards.
What does a ranked candidate look like?
Each candidate on the long list from AI candidate sourcing is scored against the full vacancy: the hard filters and the preferences. The card shows the reasoning behind the score.
- Match score: how well the profile fits the complete vacancy
- Matching evidence: which experience supports each requirement, and where it was found
- Missing requirements: what is not visible in the profile, often a question for the first call
- Risk factors: points to check, not reasons to reject
Lotte B.
- 8 yrs Javasource: profile
- Spring Boot in 3 rolessource: work history
- Payments domain since 2021source: work history
- Kubernetes not mentionedask in call
- 2 job changes in 3 yrssource: work history
How is the ranking built?
The ranking starts from the requirements set in vacancy intelligence. Hard requirements weigh most. Preferences decide the order among candidates who meet the basics. Because the criteria are fixed per vacancy, every profile is judged on the same points, whether it is the first or the three hundredth.
A missing requirement is not the same as a failed one. If a profile does not mention Kubernetes, the candidate may still have used it. The ranking shows it as missing, so the recruiter can ask, instead of silently dropping a strong candidate. You see this most in markets where people keep their profiles short, such as technical trades or healthcare.
What a risk factor is, and is not
A risk factor is a signal worth a question: several short jobs in a row, a long commute, a recent move into a different field. It is not a verdict. A string of short contracts is normal in interim work or secondment. The recruiter decides what the signal means for this role and this client.
How does this compare with resume screening software?
Many screening tools sort applications that come in through a job posting. AIcruiters ranks people it sourced itself, before anyone has applied. The logic of evidence-based scoring is similar; the moment in the funnel is different.
How do you keep a ranking fair?
Rankings should be based on predefined, job-related criteria. Bias usually enters through proxies: a gap in a CV that reflects care duties, a university name, a location that correlates with background. Keeping the criteria tied to the vacancy and the evidence visible is what lets a recruiter spot and correct a questionable outcome.
Low-confidence cases, such as conflicting work history or a profile the agent cannot interpret well, are routed to a recruiter rather than scored and forgotten. How those review points work is explained on the human-in-the-loop page.
What does the EU AI Act mean for candidate ranking?
The AI Act lists AI systems used to evaluate candidates in recruitment as high-risk (Annex III, point 4). Following the Digital Omnibus on AI (Regulation (EU) 2026/1744), the obligations for these Annex III high-risk systems apply, according to the current timeline, from 2 December 2027. They include human oversight, transparency towards users and record-keeping. Obligations fall on both providers and the organisations that use such systems.
This is not legal advice and details depend on your role and setup. Our guide to the EU AI Act in recruitment goes deeper, and the security page describes how AIcruiters approaches oversight and data control.
Questions about candidate ranking
Does a high ranking decide who gets hired?
No. A high score means the profile fits the vacancy on paper and is worth a conversation. Interviews, assessment and the final decision remain with your recruiters and hiring managers.
What if the agent ranks the wrong candidate first?
The evidence behind the score is visible, so a recruiter can see why and correct it, for example by moving a requirement or adjusting a preference. Criteria can be refined before outreach begins.
Can we decide which candidates move on to outreach?
Yes. You set approval rules for candidate selection, so a recruiter can review the ranked list before anyone is contacted. Low-confidence cases go to a recruiter either way. How far selection runs automatically for a given vacancy is agreed per implementation.
Does it screen incoming applications too?
AIcruiters is built for sourced candidates. How it fits alongside the applications in your ATS is something we look at per setup.
Can a candidate ask why they were contacted?
Because every ranking has visible evidence tied to the vacancy, the recruiter can explain which experience made the profile relevant.
Check the ranking on your own vacancy
In the demo you see the top candidates for a live role, with the evidence, gaps and risks behind every score. Challenge the ranking; that is what the evidence is for.