How the ATS Really Works: A Recruiter Who's Used 5 Explains

Job seekers are afraid of the ATS. I understand why — the phrase “applicant tracking system” sounds ominous, like a bouncer that decides if your CV even gets seen. The fear gets worse with every article online that says you must do X, Y, and Z or the ATS will auto-reject you.

Most of that content is wrong. I say this as someone who has used five different ATS platforms across 12 years of recruitment: Workday, Greenhouse, Lever, Bullhorn, and LinkedIn Recruiter. I’ve spent thousands of hours searching candidates, reviewing applications, and building filters. I know what these systems do and don’t do.

This article is the myth-busting guide I wish existed 10 years ago. By the end, you’ll understand exactly what happens when you hit “submit” on a job application, what actually gets your CV filtered out, and which parts of the ATS panic are genuinely worth addressing.

What the ATS actually is

An ATS is a database with application-specific features. That’s the honest one-sentence definition.

When you submit an application, here’s what happens:

1. Your CV is uploaded to the ATS (PDF or docx — both work)
2. The ATS parses your CV into structured fields — name, email, work experience, education, skills — using text extraction
3. Your data is indexed — meaning the ATS stores every word of your CV in a searchable format
4. Your application sits in a database — alongside everyone else who applied to the role
5. A recruiter opens the ATS and decides what to do — usually starting with a search query

That’s it. Notice what’s missing: nowhere in this flow is the ATS making a yes/no decision on your CV. The ATS is inert until a recruiter interacts with it.

The 5 biggest ATS myths, debunked

Myth 1: “The ATS auto-rejects resumes below a score threshold”

Reality: No mainstream ATS has automatic rejection enabled by default. Some have the option, but most recruiters I know don’t use it — we want to see the full candidate pool.

What people confuse with auto-rejection:

• Knockout questions on application forms (“Are you authorized to work in this country?” — “No” → auto-rejected, but this isn’t the ATS being harsh, this is a legal must-filter)
• Recruiter filters that look like auto-rejection (“Show me candidates with 5+ years experience” — filters out everyone below, but the ATS didn’t reject them, the recruiter did)
• The ‘resume score’ shown on your side — some ATS show you a match percentage; this is informational, not a rejection threshold

When someone says they were “rejected by the ATS,” 95% of the time they mean they were filtered out by a recruiter using ATS filters. It’s a human call, not a software one.

Myth 2: “PDFs don’t work — you need .docx”

Reality: Every ATS I’ve used in the past 5 years parses PDFs correctly. Workday, Greenhouse, Lever, SmartRecruiters, iCIMS — all handle PDFs. The advice to use .docx was correct in 2015 and has been wrong since about 2020.

What actually causes parsing problems:

• Multi-column CVs (two-column designs) — the ATS reads left-to-right, top-to-bottom, jumbling your content
• Text inside images (CVs designed in Canva with text baked into graphics) — the ATS can’t OCR them reliably
• Custom fonts not embedded — text shows as question marks
• Tables for layout — nested tables can break parsing

Modern PDFs avoid all of these. Use a simple one-column layout, standard fonts, save as PDF from Word/Google Docs/Notion. Done.

Myth 3: “Stuff your resume with keywords to pass the ATS”

Reality: Stuffing keywords either does nothing (the ATS ignores unnatural repetition) or actively hurts you (the bullet sounds weird to the human who eventually reads it).

Here’s how ATS keyword matching actually works under the hood: most use TF-IDF ranking or similar (basically the same algorithm as search engines). You get credit for:

• Having the keyword present (1 point, essentially)
• Having it in context (more credit)
• Having it in structured fields like “skills” or “title” (bonus credit)

You don’t get 10x credit for 10 mentions. You get diminishing returns fast, and above 5-6 mentions of the same term you look suspicious.

What to actually do: include the top 5-8 skills from the job description naturally, each mentioned 1-3 times across the CV. Don’t force it. If your job didn’t use SQL, don’t put SQL on the CV just because the JD asks for it — you’ll fail the interview.

Myth 4: “The ATS uses AI to screen you”

Reality: The ATS uses keyword matching, Boolean search, and simple filters. Not AI in any meaningful sense. This was true a decade ago and is still largely true today.

Some newer platforms are adding AI features — semantic matching, skill inference, candidate summarization. Most recruiters ignore these features or treat them as secondary signals. The primary workflow is still: recruiter types “product manager SaaS London 5 years” into a search box, ATS returns a ranked list, recruiter scrolls it.

“AI-powered ATS” is mostly marketing. The filtering that matters — whether a recruiter contacts you — is a human decision.

Myth 5: “One magic CV format passes all ATS”

Reality: Different ATS platforms parse slightly differently, but the same clean, simple CV works for all of them. There’s no magic format.

The format that consistently works:

• One column, top to bottom
• Standard section headings (Experience, Education, Skills — not “Career Highlights” or creative variants)
• Standard fonts (Arial, Calibri, Inter, Times)
• Black text on white background
• PDF export (or .docx if you prefer)
• No headers/footers for important info — some ATS strip them
• File name: “FirstName_LastName_Resume.pdf”

This isn’t ATS-specific. It’s just clean CV design. The ATS is one reason to use this format; a hiring manager scanning on a phone is another.

What the ATS actually does care about

Stripping away the myths, here’s what ATS actually does that affects you:

1. Parses your work experience into dated records

The ATS extracts each job you’ve held as a separate record with start date, end date, company, title, and description. If your CV formats dates weirdly (“2022 - Present” without commas, or “Jan ‘22-Present” with inconsistent casing), the parser may mis-associate your role dates with descriptions.

Fix: format all dates the same way: Jan 2022 – Present (with en-dash, not hyphen).

2. Extracts your skills section into structured tags

The ATS often pulls your “Skills” section into tag form — individual items like “Python”, “Figma”, “Salesforce”. Recruiters can then filter by these tags directly.

Fix: have a dedicated Skills section with 8-15 skills as a comma-separated list or bulleted one-liners. Don’t hide skills inside paragraph-form experience descriptions — they parse less reliably into tags.

3. Matches keywords in the job description

When a recruiter enables keyword matching on a role, the ATS scores each application against the JD’s text. High-match candidates show up first in their review queue.

Fix: tailor the CV to each role (see my tailoring workflow). Mirror the JD’s language where honest to do so.

4. Supports Boolean recruiter searches

A recruiter might search: "product manager" AND (SaaS OR B2B) AND "London" AND "5 years". The ATS checks which applications match all clauses and surfaces them.

Fix: include obvious variants of your skills. If you do “UX research,” include the full phrase and “user research” as synonyms in different places. Same idea. The recruiter’s Boolean search string is what’s running under the hood, and synonyms are how you stay in the result set. For career changers, this synonym layer is doubly important because the target industry’s vocabulary will not match your old one out of the box.

5. Stores your application indefinitely

Unlike spray-and-pray inbox approaches, your ATS application persists. When I search for new roles later, your CV shows up if you match. This is why keeping a baseline CV with all your experience (for later searches) matters.

Fix: the baseline version of your CV should be comprehensive. You tailor specific applications, but the “in the ATS” version stays rich so recruiters find you for future roles too.

The 5 real reasons CVs get filtered out

Forget ATS score. Here’s what actually causes rejection, based on what I’ve seen in my queues:

1. Mismatch on must-haves (years of experience, location, work authorization). The ATS doesn’t reject you — the recruiter filters by these criteria and you fall out.
2. No clear narrative in the summary — recruiters spend 6-8 seconds on initial scan. If your summary is “Results-driven professional leveraging…” we move on.
3. No measurable outcomes — bullets like “Responsible for managing projects” with no numbers or specifics get skimmed past.
4. Irrelevant experience emphasized — your most relevant role is buried in page 2, while a 2015 side project gets top billing.
5. Writing that screams AI — see the 13 buzzwords article. If I spot four AI tells in paragraph one, I move on.

None of these are “ATS problems.” They’re human-recruiter problems. The ATS is a tool we use, not the gatekeeper.

How to “optimize for ATS” (the real version)

Given all of the above, here’s what actually helps:

✅ Do:

• Keep the format simple (one column, standard headings, PDF)
• Tailor keywords per role from the job description
• Put your most relevant experience first
• Use consistent date formatting
• Have a dedicated Skills section (8-15 items)
• Save as FirstName_LastName_Resume.pdf

❌ Don’t:

• Use two-column “creative” templates
• Stuff the same keyword 10+ times
• Rely on tools that promise ATS-beating magic
• Buy into the “score above 80%” hype from resume checkers
• Add a “keyword stuffing” white-text section at the bottom (this gets you blacklisted if noticed)

Where AI tools fit in

Dedicated tools can genuinely help with the 5 real rejection reasons:

For keyword matching to the JD: Rezi shows your resume-to-JD match percentage in real time, and highlights missing keywords. Useful for confirming you’ve covered the ground. Don’t chase a specific score — use it as a checklist.

For application tracking: Teal saves each job description you view, tracks which CVs you sent where, and stores your tailored versions. Keeps you organized across 20+ applications.

Neither tool “beats the ATS” because there’s nothing to beat. Both help you write better tailored CVs, faster. That’s what matters. For the detailed breakdown of which tool fits your situation, see Teal vs Rezi.

What I tell candidates about the ATS

“Stop worrying about the ATS. Write a clean, specific, tailored CV and it’ll pass through any modern ATS fine. The recruiter on the other end is the one you need to impress, and that’s about content — specific outcomes, clear writing, tailored relevance. The ATS is a neutral tool. The human is the filter.”

Three related articles to deepen this:

• How to tailor your resume to a job description with AI — the 15-minute workflow
• 13 AI resume buzzwords recruiters hate — what makes your CV look generic
• ChatGPT prompts for resume writing — the prompts I actually recommend

ATS panic costs job seekers thousands of hours of overthinking and drives them toward gimmicky tools. Skip it. Write clear, tailored, specific CVs. The system is not your enemy.

Related reading

• UK CV format 2026 — the 7-section UK CV that survives the 8-second skim and parses cleanly through every ATS I’ve used.
• Why am I not getting hired? — 11 specific recruiter rejection reasons that have nothing to do with the ATS.
• LinkedIn Easy Apply: why it fails — the ATS queue on the LinkedIn side, and how to route around it.
• LinkedIn skills to add in 2026 — the skill-filter layer that works parallel to CV keyword parsing.
• LinkedIn Open to Work: what recruiters see — the search-visibility flag that sits above the ATS.
• Cover letter length: the recruiter word-count test — the document humans read after the ATS has done its thing.

Tools that help

• Jobscan review — the ATS scanner I recommend for diagnosing keyword gaps. Free tier is enough for most candidates.
• Rezi review — the ATS-first CV builder if you need to rewrite, not just score.