Can an ATS screen resumes? What it does and does not do
Oct 5, 2026 · 7 min read
Can an ATS automatically screen resumes? Yes, partly. An applicant tracking system stores applications and can filter them using rules you set, and some also rank or score them. What it does not do by default is understand a CV the way a recruiter does, so a person has to stay involved.
What does an ATS actually do?
An ATS is a system of record for hiring. It collects applications, stores candidate details, moves candidates through stages and keeps a history of who did what. Many include job posting, email templates and basic reporting. Screening is one of several features, and it varies a lot between products.
- Collects applications from job boards, a career page and email in one place.
- Parses CVs into fields such as name, past employers and skills, with varying accuracy.
- Moves candidates through stages that your team defines.
- Lets you set knockout questions and filters, such as work authorisation or location.
- Records notes, status and communication so the team shares one view.
How is a keyword filter different from scoring against a job description?
A keyword filter checks whether certain words appear in a CV and passes or fails the candidate. Scoring against a job description compares the whole CV with the requirements of the role and gives a number or ranking, usually with some reasoning. The first is simple and strict. The second is more flexible but needs checking.
| Approach | How it works | Strength | Weakness |
|---|---|---|---|
| Keyword filter | Passes a CV if it contains chosen words or phrases | Easy to explain and consistent | Misses people who describe the same skill in different words, and passes people who only list a word |
| Knockout question | Asks a yes or no question on the application form | Handles hard requirements such as location or a licence | Only as good as the question and the candidate's honesty |
| Scoring against a job description | Compares the whole CV with the requirements and gives a score | Can recognise similar skills and relevant experience | Needs checking, and the score depends on how clear the job description is |
| Manual review | A person reads the CV | Understands context | Slow when there are many applicants, and varies between readers |
Neither automatic approach is complete. The right choice depends on how many applications you receive and how clear your requirements are.
What does ATS score mean?
ATS score is a loose term, and it does not have one standard meaning. Sometimes it means how well a CV matches a job description in a particular tool. Sometimes it means how easily a system can read the CV file. There is no universal score that every ATS uses, so a number from one tool cannot be compared with a number from another.
The useful question is what the score is based on. A score that comes with a written reason, such as which requirements the CV met and which it did not, can be checked. A bare number cannot.
Can AI resume screening software do better than keywords?
It can recognise related skills and relevant experience that a keyword filter would miss, for example a candidate who describes managing a support queue without using the exact job title. It can also read the full CV for context. It can still make mistakes, so its output should be treated as a ranked starting point for a person to check.
AI screening agents for an ATS are tools that sit beside or inside an applicant tracking system and score or rank applications. Ask how they work, what they are given as the requirements, and whether you can see the reasons for each score.
How do you check a resume screening tool?
Test it on real CVs you already understand, and ask for the reasons behind every score. Do this before using it on a live role. A tool that cannot explain its scores, or that gives surprising results on CVs you know well, needs more caution.
- Ask for the reasons behind a score, in words, tied to the CV and the job description.
- Test with known CVs. Include a clear strong match, a clear mismatch and a few borderline cases from past hires or past rejections.
- Include a CV from someone with an unusual background and see whether the score is fair.
- Try a CV that uses different words for the same skill and see whether the tool still recognises it.
- Change the job description slightly and see whether the scores change in a sensible way.
- Ask whether anything rejects candidates automatically, and whether you can turn that off.
What does a good test look like?
A good test uses a small set of CVs where you already know the answer, and compares the tool with your own judgment. Write down your view of each CV before running the tool, so you do not adjust your opinion afterwards.
| Test CV (placeholder) | Your view first | What to check in the tool |
|---|---|---|
| [CV 1], clear strong match | Should be near the top | Is it ranked high, and do the reasons mention the right requirements |
| [CV 2], clear mismatch | Should be near the bottom | Is it ranked low for the right reasons |
| [CV 3], strong skills, different wording | Should be shortlisted | Does the tool recognise the skills despite the wording |
| [CV 4], keyword heavy, little real experience | Should not be near the top | Does the tool see through a list of terms |
| [CV 5], career break or unusual path | A person would want to read it | Is it treated fairly, and is it flagged for review and not dropped |
Where must a person stay involved?
A person should read the evidence for every shortlist decision and make every rejection decision. Automatic screening is best used to sort a large pile into an order, so your team reads the most relevant CVs first. It should not be used to throw away applications that nobody has seen.
- A person reviews the reasons behind each score before shortlisting.
- A person looks at a sample of low-scored CVs to check they were scored fairly.
- A person decides whether to reject, and the tool does not do it on its own.
- A person keeps the job description clear and up to date, because scores depend on it.
Where AI Interview Agents fits
AI Interview Agents offers resume screening. You can upload up to 500 CVs in one batch, and each is scored from 0 to 100 against the job description. Shortlisted candidates can be sent to an AI interview, which is a two-way spoken first-round interview in the candidate's browser, with a transcript and a report for your team.
It also offers agentic resume search over your own talent pool, returning candidates with verbatim CV evidence, so you can see the exact words from the CV that support each match. A person on your hiring team makes every hiring decision. AIIA scores and reports, and it does not reject candidates on its own. You can start on the free plan and test it on CVs you already know.