Blog/AI In HR

CV screening: what it is, a worked example and 5 steps

Flowmingo Editorial TeamFlowmingo Editorial Team5 minut čtení
CV screening: 5 fair steps, with a worked example: CV screening on 1 support lead role: 244 CVs arrive, 146 pass the must-haves, 19 show evidence and go to interview

CV screening is the first check of each applicant's CV against a job's must-have requirements. It decides who moves on to an interview.

You meet CV screening, which US recruiters call resume screening, the moment a role goes live. It is usually the step that decides who gets an interview, and on Greenhouse, an applicant tracking system, the average North American job drew more than twice as many CVs in 2025 as in 2022.

The figures come from Greenhouse's North American benchmark, which draws on 640 million applications made through its platform from 2022 to 2025. Applications per job went from about 115 in 2022 to 244 in 2025.

With that many CVs, what keeps the screen both fast and fair is deciding what you are looking for before you open the first one.

Key takeaways

  • What does CV screening decide? Who gets an interview. In Greenhouse's 2025 North American figures, 7.6% of applicants passed the first stage, about 19 of every 244.
  • Is resume screening the same thing? Yes, under another name: "resume" is the US word, and "CV" is the one used in the UK and Asia.
  • What should a CV screen look for? Proof of the job's must-have skills, not years. In a 2019 review of 11,785 workers, past work experience told almost nothing about how well they later did the job, a finding now more than 6 years old.
  • Can AI screen CVs? It can score them, but test it first. In one 2026 study, 22 AI models shown pairs of equally strong CVs picked the first one 63.5% of the time.

1. What is CV screening in recruitment?

CV screening is the step where someone reads each CV against the job's criteria and sorts the applicants into yes, no and maybe. It sits after the application and before any interview.

The screen runs 2 kinds of check. Knockout checks are pass-or-fail facts, such as the right to work or a required licence. Evidence checks need a closer read, because they ask whether the CV shows the person has done this kind of work, with results.

It also helps to know what CV screening is not. It is not resume parsing, which only copies a CV into the fields of an applicant tracking system. Nor is it an interview, because a CV shows what a person claims, not how they think.

That difference sets the rule for the whole screen: cut only on what a CV can prove.

2. What does CV screening look like on a real role?

On a real role, the screen works like a funnel. Knockout checks remove the clear no's, an evidence check ranks the rest, and the top slice goes to interview.

The role below is an example built around 2 real figures. The 244 CVs and the 7.6% first-stage pass rate are 2025 averages from Greenhouse, which does not define its stages. The split in between, 98 and 127, is our illustration.

Say you are hiring a customer support lead, and 244 CVs arrive in a week. Your must-haves are facts a CV either shows or does not, such as the right to work in Singapore and written English.

Time spent in support is deliberately not on that list, because, as step 3 below explains, years in a job say little about later performance. The evidence check looks for proof of the skill instead: a team or queue the person ran, and a result they can name.

Resume screening example for a support lead role: 244 CVs, 98 fail a must-have, 127 show no evidence and 19 go to interview

CV screening on 1 support lead role: 244 CVs arrive, 98 fail a must-have, 127 show no evidence and 19 go to interview.

The 98 applicants who miss a knockout leave first, and that takes minutes, because each check is a fact. The 146 who remain are read for evidence, and 127 of them never mention a queue, a team or a result.

That leaves 19 people, which is what the Greenhouse benchmark's 7.6% first-stage pass rate gives on 244 CVs, rounded to the nearest person. Those 19 are the ones you invite to interview.

3. How do you screen CVs fairly, step by step?

Write the criteria before you open the first CV, then apply the same checks, in the same order, to every applicant. That keeps the screen fast and lets you explain every no.

  1. Write 3 to 5 must-haves and 2 to 3 evidence criteria from the job. Singapore's Tripartite Alliance for Fair and Progressive Employment Practices (TAFEP) tells employers to list selection criteria and apply them consistently to all candidates.
  2. Run the knockout checks first. They are facts, so they are quick to check and hard to argue with.
  3. Look for proof of the skill, not years. In a 2019 review, now more than 6 years old, Van Iddekinge and colleagues pooled studies of 11,785 workers. Prior experience tracked their job performance at only 0.06, where 0 means no link at all and 1 a perfect one.
  4. Score each CV on the same criteria with the same rule. That way a score means the same thing on every CV, so you can rank them and invite the top slice your interview hours can hold.
  5. Write 1 line on why each person moved on or stopped. That line is what lets you answer a candidate or a hiring manager who asks later.

3.1 Which mistakes make CV screening unfair?

A screen turns unfair when something other than your criteria changes the result. As 2 studies show, that can happen through the name on the CV and through the order in which CVs are read.

In a 2022 study in the Quarterly Journal of Economics, now more than 4 years old, Kline, Rose and Walters sent more than 83,000 test applications to 108 of the largest US employers.

Distinctively Black names cut the chance of employer contact by 2.1 percentage points, compared with distinctively white names. For a screen like the one above, that is the reason to let the criteria you wrote, not the name at the top of the CV, decide each yes and no.

The order of reading can tilt the result too, and here the evidence comes from software. In a 2026 study of 22 AI models given pairs of equally strong CVs, the first CV shown won 63.5% of the time (Rozado, PeerJ Computer Science).

Since the 2 CVs in each pair were equally strong, the first should have won only about half the time. So shuffle the order, or check that your tool scores each CV on its own.

With names and order kept out of the scores, what remains is deciding how many people to invite, and our guide to candidate shortlisting shows how to set that cut line.

4. How do you run CV screening with Flowmingo?

You set the criteria once, and every CV gets a score against them, with the reason written next to each score. You then read the evidence and decide who moves on.

Before you start, have 2 things ready: the job description and the 3 to 5 must-haves from step 1.

A tool can only score what your criteria name, so write each one as something a CV can show. For example, "has handled a live chat queue" gives a clear score, while "great communicator" leaves the scorer to guess.

Keep writing the 1-line reason from step 5 as well, so the hiring manager can see why each person moved on or stopped.

How Flowmingo helps

Paste your job description, check the must-haves, and Flowmingo scores every CV against them, rating each requirement 0 to 10 and weighting it by importance.

You get a ranked shortlist with the reasoning attached, and the decision stays with your team. Sign up free to run your next CV screening on Flowmingo.

5. Sources

Every study, benchmark and guideline in this glossary entry links to its source below.

  • Greenhouse. (2026, March). The Hire Standard: Greenhouse benchmark report, North America. Greenhouse
  • Kline, P., Rose, E., & Walters, C. (2022). Systemic discrimination among large U.S. employers. Quarterly Journal of Economics, 137(4), 1963-2036. NBER w29053
  • Rozado, D. (2026). Gender and positional biases in LLM-based hiring decisions: Evidence from comparative CV/resume evaluations. PeerJ Computer Science, 12, e3628. PeerJ
  • Tripartite Alliance for Fair and Progressive Employment Practices. (n.d.). Conducting job interviews. Retrieved 26 October 2026. TAFEP
  • Van Iddekinge, C. H., Arnold, J. D., Frieder, R. E., & Roth, P. L. (2019). A meta-analysis of the criterion-related validity of prehire work experience. Personnel Psychology, 72(4), 571-598. UNF Digital Commons

Zaregistrujte se a začněte za cca 60 sekund. Bez karty, bez hovoru

Zaregistrujte se zdarma
Flowmingo Editorial Team

Flowmingo Editorial Team

We write practical guides for recruiters and HR teams who want to hire faster and more fairly. Each guide draws on hiring research, employment rules and Flowmingo's own data from real interviews, and lists its sources.

LinkedInYouTubeFacebook