What an ATS Does to Your Resume: Parsing, Knockouts and Search

What is an applicant tracking system?

An ATS is the recruiting team's database and workflow tool. Every open job gets a record, every application is attached to it, and every candidate moves through stages such as "new", "screening", "interview" and "offer". Recruiters use it to post jobs, email candidates, schedule interviews, collect feedback from hiring managers and report on hiring. Reading resumes is only one part of what it does.

You have almost certainly used one. Workday is a large HR and finance suite whose recruiting module is common at big enterprises. Greenhouse and Lever are recruiting platforms popular with tech companies and scale-ups. iCIMS is a talent acquisition platform used by many large employers. Taleo is Oracle's long-established recruiting system, and SmartRecruiters is a hiring platform used by mid-size and large companies. Many European employers also use regional systems such as Teamtailor. The interfaces differ, but the basic pipeline is the same: collect, parse, screen, search.

What happens when you click "Apply"?

An application usually has two parts: a form and a file. The form asks for contact details, sometimes your work history again, and a set of screening questions. The file is your resume, plus a cover letter if one is requested. Many systems read the uploaded resume first and use it to pre-fill the form — which is the one place you get to see parsing in action. If the pre-filled job titles, employers or dates come out wrong, that is roughly what the system has stored about you. Correct the fields by hand before you submit.

Once submitted, your application lands in the pipeline for that job with the parsed data, your answers and the original file attached. A recruiter can open the original document, but the version of you that can be searched and filtered is the parsed one.

How does resume parsing work, and what breaks it?

Parsing turns a document into data. The parser extracts the text layer from your PDF or Word file, works out the reading order, splits the text into sections by recognising headings, and maps pieces of text to fields: this line is a job title, this is an employer, these are dates, these words are skills. Every step can go wrong, and errors compound — a scrambled reading order produces wrong sections, which produce wrong fields.

The failure patterns are well known:

• No text layer. A scanned resume, an image export or a design tool that turns text into shapes gives the parser nothing to read. Test it: open your PDF, select all, and paste into a plain text editor. If clean text doesn't come out, a parser can't read it either.
• Columns. Parsers read in a sequence, and a sidebar can get interleaved with your job history, so a skills list ends up inside a job description. Modern parsers handle well-built two-column layouts better than they used to, but single-column is still the most predictable.
• Headers and footers. Some parsers skip them, so contact details placed there can vanish from your record.
• Tables and text boxes. Content inside them may be read out of order, merged together or dropped.
• Graphics. Skill bars, star ratings, icons used instead of labels, and charts carry no text. "Python" next to four filled dots becomes just "Python" — or nothing.
• Creative headings and inconsistent dates. "Where I've made an impact" doesn't map to any section, and a mix of "2021", "03/2022" and "Spring 2023" makes it hard to calculate how long you held each role.

None of this means your resume has to look plain. It means structure should come from real text, standard headings and a simple layout — not from boxes and images. The full rule set is in our ATS-friendly resume checklist.

What are knockout questions?

Knockout questions — also called screening or pre-qualifying questions — are the form questions an employer configures for a specific job. Are you authorised to work in this country? Do you hold the required licence? Can you work on site in this city? Do you have at least three years of experience with this tool? What are your salary expectations?

This is where genuinely automatic filtering usually happens, and it runs on your answers, not your resume. Depending on how the employer has set up the job, an answer outside the required range can move your application straight to a rejected status or to the bottom of the list. Two practical consequences: read every question slowly, because a misclick on a yes/no question can end an application you were qualified for. And answer honestly — overstating a requirement might get you past the form, but it tends to come out in the first conversation.

How do recruiters search and rank candidates?

For an open role, a recruiter typically works through the applications that came in for it, sorted and filtered by stage, knockout answers, location or date. With a large pool, or when looking back through past applicants, they search. That search is usually keyword-based, sometimes with Boolean operators: "product manager" AND (fintech OR payments). If those terms are not in your parsed text, you do not appear in the results, however qualified you are.

Many platforms now also offer match scores, ranking or AI-assisted recommendations. What these features do, and whether an employer has switched them on, varies a lot between vendors and companies, so you can't optimise for one specific algorithm. What holds across all of them is simple: the words of the job posting, used honestly in your resume, help both a keyword search and a semantic match. Our guide to mirroring a job description without stuffing shows how to do that and still sound like a person.

Does an ATS automatically reject most resumes?

You have probably seen the claim that 75% of resumes are rejected by software before a human sees them. We have not found a credible, traceable source for that number, and it does not match how these systems are built: an ATS is primarily a database with search and workflow on top, and the reject button is usually pressed by a person.

That does not mean nothing is automated. These are the things that realistically keep an application from being seen:

• A knockout answer the employer set up to disqualify.
• A parse so broken that your experience does not show up in searches or filters — invisible rather than rejected.
• Missing the language of the posting, so you sit low in a keyword search or a match view.
• Timing: a popular role can move to interviews before every application has been read.
• A recruiter's judgement after a quick scan — covered in how recruiters read your resume.

The takeaway is the same whichever way you look at it: make your resume easy to parse, use the posting's real language, and answer screening questions carefully.

How does KaizenCV's ATS score work?

No outside tool can reproduce a specific employer's ATS — their parser settings, knockout rules and recruiter searches are private. So the ATS score in KaizenCV's resume builder doesn't pretend to be Workday or Greenhouse. It checks the things every system and every recruiter depend on, and it shows its working:

• Format and content checks run in code, not through an AI model: contact details, standard section headings, consistent and possible dates, leftover placeholder text, length, bullet points, quantified results, action verbs, duplicated lines and keyword stuffing. The same resume always gets the same result.
• When you add a job posting, AI reads it once and turns it into a fixed checklist of requirements, each marked as must-have, strong or nice-to-have and tied to the posting's own words. A requirement whose quoted wording can't be found in the posting is dropped, so the checklist can't ask for something the employer never did.
• Matching your resume against that checklist happens in code, using the requirement's wording plus known aliases and synonyms. An optional deeper check uses AI to recognise when you describe a requirement in different words.
• The score is shown as points earned out of the points checked, and each finding shows how many points fixing it is worth. Serious problems — no email address, an impossible date range, placeholder text, or a personal identity number such as a social security number — are flagged as blockers rather than buried inside the number.

Want a quick read before building anything? The free ATS score checker is a separate, AI-generated report: upload your resume, add a job description if you have one, and get a score, a parseability breakdown, a keyword match and a prioritised list of fixes — no account needed.

How do you make your resume ATS-friendly?

• Start from a template that exports real, selectable text. Every one of the KaizenCV resume templates does; pick a single-column design if you are applying through portals you don't know.
• Keep contact details in the main body of the page, as plain text.
• Use standard headings: Experience, Education, Skills.
• Write every date the same way, with month and year.
• Replace skill bars and icons with words.
• Mirror the job title and must-have skills from the posting — wherever they are true for you.
• Run the copy-paste test on your exported PDF, and check the pre-filled application form before you submit.
• Answer every knockout question slowly and honestly.

Do those eight things and the ATS stops being the obstacle. What is left is what should decide the job anyway: whether your experience fits the role, and whether your resume makes that obvious in a few seconds.