Resumes and ATS in 2026: the complete guide to getting past the filters

How an ATS actually reads your resume, and how to make yours compatible: keywords, section headings, file format, the mistakes to avoid, and a free scan.

You apply to roles that match your profile, your resume is well written, and still nothing comes back. Before blaming your experience, ask a different question: was your resume ever read properly? Between the moment you click "Apply" and the moment a recruiter looks at anything, there is almost always a piece of software — an applicant tracking system — that first turned your document into plain text, split it into sections and compared it to the job posting.

This guide explains what an ATS actually does, what breaks the reading of your resume, which keywords to reuse, and how to check the result before you send. At the end you can test your resume for free against a real posting and see exactly what is missing.

We also checked the one thing that does reject people on its own. On 2 October 2026 we read the application forms of 46 US job postings at 10 companies that publish their Greenhouse boards. 45 of them would not let you submit without answering a work-authorization or visa-sponsorship question. That form, not your resume layout, is where most automatic rejections come from.

What an applicant tracking system actually is

An ATS is the tool where resumes land when you apply through a careers site, a job board, or a platform like Workday, Taleo, SmartRecruiters, Lever or Greenhouse. Recruiters use it to centralise applications, search them, annotate them and move each candidate from one status to the next.

The part that matters to you: an ATS does not look at your resume the way a person does. It builds a text version of it, tries to rebuild your profile field by field, then compares that profile to the requirements of the role. Anything it cannot extract does not exist as far as it is concerned.

Three steps: extraction, matching, scoring

Nearly every ATS runs the same three operations.

1. Extraction (parsing). The software opens your file and converts it to plain text. It then tries to recognise blocks: name, email, phone, experience, education, skills. This is the fragile step. Two columns, a table or an unusual heading is enough to scramble the blocks: a job ends up filed as education, a date attaches to the wrong employer, a skill disappears.

2. Matching. The ATS compares the extracted text to the terms in the posting: job title, tools, degrees, languages, years of experience, location. The search is often literal. "Project management" and "project delivery" are not always treated as equivalent, and a system looking for "Power BI" will not necessarily find "PowerBI".

3. Scoring and ranking. The recruiter then sees an ordered list: a match score, or simply the result of a keyword search ("Python" AND "SQL" AND "fluent English"). If a posting draws 300 applicants, that ranking decides what gets opened and what never does.

In other words, your resume has three doors to get through: be readable, be recognised as relevant, then rank high enough to be opened.

How many employers actually use one

This is where the internet is least reliable. The famous "75% of resumes are never seen by a human" has been repeated for a decade with no traceable source behind it, and it has quietly become the sales pitch of every tool selling ATS anxiety.

The honest picture is that adoption depends almost entirely on company size.

Adoption data outside France is patchier and varies by market, but the size effect holds everywhere: a 30-person business will read your resume straight out of an inbox; a large group, a consultancy, a bank or a recruitment agency will almost certainly run it through an ATS first.

What that changes for your strategy:

• Targeting large employers, agencies, or anything posted on a corporate careers site? Assume an ATS is in the loop, always.
• Applying to small companies, startups, or through a referral? Your resume often reaches a human directly — though nothing stops it being loaded into a system later.
• Either way, a resume a machine can read is a resume a busy recruiter can skim. Optimising for the ATS costs the human reader nothing; it helps them.

The myth worth killing: an ATS ranks, it does not reject

This is the most common misunderstanding, and the one the fear-selling tools depend on: "the robot throws your resume in the bin." In the overwhelming majority of cases, that is simply false.

An ATS is almost never allowed to delete an application. It sorts, ranks and filters what is displayed. Your resume stays in the database; it is just on page 4 of a list the recruiter will not scroll to the end of. The practical result feels like rejection, but the mechanism is different — and that difference is exactly what makes optimisation worth doing.

Two caveats:

• Some application forms include knock-out questions: work authorisation, driving licence, minimum years of experience, willingness to relocate. There the rejection genuinely is automatic — and it comes from your answers in the form, not from your resume's layout.
• If extraction fails completely — a resume sent as an image, text that cannot be selected — your candidate record is close to empty. You are not rejected: you are invisible to every search the recruiter runs, which amounts to the same thing.

The conclusion is more encouraging than it sounds: you are not playing against an algorithm that eliminates you, you are playing for a position in a ranking. Every accurate keyword and every conventional heading moves you up the list.

Recruiters say the same thing when asked. Enhancv interviewed 25 US recruiters using more than ten different systems, and 23 of them said their ATS does not reject resumes for formatting, content or design. The 75 % figure has no study behind it. It goes back to an unpublished study by a resume-writing company, around 2012, that no longer exists.

The knock-out questions, counted on real forms

Every page on this topic says "watch out for knock-out questions". None of them says how common they are, so we counted.

The method takes ten minutes to reproduce. Greenhouse, the most common ATS in Huntr's 2026 report on the jobs its users saved, exposes every public job board through an open API, and adding ?questions=true to a job returns the application form with each question and whether it is required. On 2 October 2026 we pulled up to five US postings from each of ten employers with public boards (Airbnb, Stripe, Discord, Dropbox, Figma, Reddit, Robinhood, Pinterest, GitLab, Cloudflare), 46 forms in all, and read every required question.

• 45 of 46 required an answer to a work-authorization or sponsorship question ("Are you legally authorized to work in the United States?", "Will you now or in the future require sponsorship?"). The only exception was one Robinhood engineering posting.
• 32 of 46 required your location or your agreement to work from a named office: a country, a state, a zip code, "are you based within 50 miles of the office?".
• 0 of 46 asked for a salary expectation as a required field.

A required question is not proof of an automatic rejection. Whether a "yes, I need sponsorship" sends you straight to the rejected pile is a rule the employer configures, and you cannot see it from outside. But it is the only place in the process where a rule can reject you without anyone reading anything, and it runs on your answers, not on your resume.

What to do with that:

• Answer these questions truthfully and read them twice. "Will you now or in the future require sponsorship?" catches people on a temporary visa who answer about today.
• If you are close to a location limit (60 miles instead of 50, a move already planned), say so in the free-text field when there is one rather than ticking "no".
• If a form's hard requirement rules you out, a referral or a direct message to the hiring manager is a better use of your time than a second application.

Is AI in the ATS rejecting you?

Some systems now add a match score or a grade to each application: Huntr's report describes four-tier match buckets in Greenhouse and A to D grades in Workday. In those tools the grade sorts the list and a person still clicks "reject". The legal risk sits with the employer when software does more than sort. Huntr notes that a US court let Mobley v. Workday, an age-discrimination case over Workday's screening tools, proceed as a nationwide collective action in May 2025.

For you, the practical rule is unchanged. A grade is computed from the same extracted text as everything else, so a resume the parser reads in order, with the posting's own words in it, is also the resume that grades well.

What the law lets you ask

None of the ranking guides mention this, and it is the most useful thing to know if you think software decided for you.

• Applying in the EU. Article 22 of the GDPR says: "The data subject shall have the right not to be subject to a decision based solely on automated processing, including profiling, which produces legal effects concerning him or her or similarly significantly affects him or her." Where an exception applies, the employer must still offer "at least the right to obtain human intervention on the part of the controller, to express his or her point of view and to contest the decision" (Article 22). Article 15(1)(h) lets you ask whether such a decision was made and for "meaningful information about the logic involved" (Article 15). A ranking that a recruiter then reads is not a decision "based solely" on automation. An automatic rejection from a knock-out rule can be.
• Applying in New York City. Local Law 144 requires an employer that uses an automated employment decision tool to have it bias-audited within the past year, to publish the audit results, and to give candidates notice "10 business days prior to use" (NYC Department of Consumer and Worker Protection). If you got that notice, the audit summary should be on the employer's website.

A short, polite email to the employer's privacy contact asking whether your application was rejected by an automated decision costs nothing.

Reuse the posting's keywords, word for word

The whole of ATS optimisation fits in one sentence: the vocabulary of your resume should be the vocabulary of the posting.

Pull the right terms out of the ad

Take the job you are targeting and note, exactly as they are written:

1. The exact job title ("Talent Acquisition Partner", "Back-end Developer", "Financial Controller"). Use it as your resume headline when it genuinely describes what you do.
2. The named tools and technologies: SAP, Salesforce, Excel, Python, Figma, Jira, Power BI.
3. The methods and domain skills: Agile, Scrum, GDPR, IFRS, lean manufacturing, marketing automation.
4. Certifications, degrees and languages: PMP, CFA, BSc, "professional working proficiency".
5. Anything the requirements section repeats two or three times: that is the core of the need. Repeats in the company pitch or the footer do not count. In 93 postings we counted, the term an ad repeats most was missing from its requirements in 29.

To make that pass faster, paste the requirements section into our free job description keyword extractor: it sorts the skills, tools, degrees and soft skills it finds and counts how often each one comes back.

Put those terms in the right place

A keyword stranded in a skills list persuades nobody; the same keyword inside a dated, quantified achievement persuades the software and the recruiter.

A few simple rules:

• Match the posting's spelling, then add the variant in brackets where one exists: "SEO (search engine optimisation)". You cover both possible searches.
• Spell out the acronym and its long form at least once: "CDP (Customer Data Platform)".
• Put the key terms in the headline, the summary and the two most recent roles — that is what a recruiter reads first once the record is open.
• Don't split a key phrase with a number: "account portfolio (40 clients)" is found by a search for "account portfolio"; "portfolio of 40 accounts" may not be.
• Tailor the resume per posting. Keywords change from one ad to the next, even under the same job title, so a generic resume misses some of the terms the recruiter will search for.

What not to do

The old workarounds are now actively harmful: white keywords on a white background, terms hidden behind an image, a wall of sixty listed skills. Hidden text is still text: it shows up in plain view in the extracted version the recruiter reads, and a recruiter who spots it drops the application, that time for good. Only claim what you can do: the bluff lasts until the first technical interview.

Use standard section headings

An ATS recognises your sections by their titles. A creative heading is a heading it fails to file.

An order that works in every ATS:

1. Contact details — first name, last name, email, phone, city, LinkedIn URL. In the body of the document, never in the file header.
2. Headline — the title of the role you are targeting, taken from the posting.
3. Summary — three lines: your function, your years of experience, two or three key skills from the ad.
4. Work experience — reverse chronological. For each role: title, company, city, dates as March 2022 – Present, then two to four achievement bullets with a number in them.
5. Education — degree, institution, year.
6. Skills — tools and technologies, as plain comma-separated text.
7. Languages and certifications.

Two details that cost more than they look:

• Dates. Write month and year on both sides (January 2021 – March 2024). An ATS computes your years of experience from those dates; "2021-2024" or "3 years" is misread when it is not ignored outright.
• Contact details. Many systems read neither the header nor the footer of a document. An email placed there can vanish from the candidate record entirely.

If the substance needs work too, our guide to writing an effective resume covers achievements, numbers and the hierarchy of information; this page covers the technical side.

The headings change with the country, not the rule: applying in France means French section names and French keywords, which our guide to the French CV format lists side by side with their English equivalents.

Choose the right file format

This is where most resumes are lost, and it is also the easiest thing to fix.

PDF or Word?

.docx is the safest format: it is the one every parser was written against, and the only one some older systems handle without degradation.

A PDF is perfectly fine, on one condition: its text must be selectable. If nothing selects with the mouse, your resume is an image and the software will read precisely nothing from it.

Selectable is not the same as readable, though. Our own PDF extractor regularly hits three cases where copy-paste "works" and the result is still unusable:

• Two columns: extraction follows the file's internal order, not reading order, and interleaves the sidebar with your experience line by line ("Skills Project manager at… Excel English 2021–2024…").
• Spaced-out letters: some templates place every letter separately, so the text comes out as "J a n e D o e". There are characters, but no words.
• A font with no character map: the text comes out as symbol soup or codes like (cid:42), including when you paste it elsewhere.

The real test takes thirty seconds: select all of the resume, paste it into a plain-text editor (Notepad, TextEdit in plain-text mode), and read it back. Whole words, in the order a person would read them? Your file passes. If not, re-export it from a single-column template.

Why do columns and sidebars fail so badly? Because a PDF stores text in drawing order, not reading order — flattened, two columns interleave line by line and dates detach from their roles. We ran three layouts of the same resume through our own extractor and published the raw output in our Canva resume test; the scrambled text is worth seeing once.

An ATS-readable resume is not an ugly resume. It is a single-column, well-spaced document with a clear hierarchy and some colour in the headings. The constraint is on structure, not on craft.

Test your resume before you apply

You can follow every rule on this page and still miss two decisive keywords. The only way to know is to run the test: convert your resume to text, compare it to the posting, and look at what is missing. Several tools do this — we compared the free ATS resume checkers honestly, ours included.

That is exactly what HiredFolio's free ATS scan does: drop in your resume as a PDF, paste the posting you are targeting, and within seconds you get a score out of 100 with the detail behind it — required skills absent from your resume, contact details the parser could not find, experience entries with no bullets, a resume that is too short or too long. No account, no card.

Test my resume for free

Once the gaps are visible, HiredFolio helps you close them in the same session: an editor with single-column templates that are already compatible, AI rewriting for your experience bullets, and a record of which version you sent for each application — useful the day a recruiter calls and you need to know which resume is in front of them. Our article on the job application tracker covers that part.

What is a good ATS score?

There is no such thing as one ATS score. The number a checker gives you, ours included, is that tool's own formula. The recruiter's system computes its own grade, if it computes one at all, and never sees yours. Two checkers can give the same resume very different numbers.

Here is ours, so you know what you are reading. The score blends four parts: keyword match against the posting (40 %), coverage of the skills the posting lists as required (20 %), formatting and extraction checks (15 %), and how closely your experience fits the role, judged by a language model (25 %). When a part has nothing to measure, for example a posting that lists no required skills, its weight moves to the others instead of counting as zero.

So do not chase a number. Read the list of missing keywords and required skills, add the ones that are true for you, and stop when the remaining gaps are things you genuinely lack.

Your checklist before you send

Three minutes per application, in this order:

• Pasted into a plain-text editor, my resume reads in order, in whole words.
• My resume is a single column, with no tables or text boxes.
• My contact details are in the body of the document, not the header.
• My headline repeats the exact job title from the posting.
• The tools and methods named in the ad appear, spelled the way the ad spells them.
• My sections use standard names (Work Experience, Education, Skills).
• My dates have a month and a year, on both sides.
• No information is trapped inside an image or a gauge.
• The file is named firstname-lastname-role.pdf.
• I answered the form's eligibility questions carefully, including "now or in the future" on sponsorship.
• I ran the scan against the posting and fixed what came back.

And remember the resume is only one link in the chain: your LinkedIn profile is indexed by the same recruiter keyword searches, and your cover letter usually lands in the same candidate record.

Read next

• How to write an effective resume in 2026: structure, examples and mistakes to avoid

  Build a resume that is clear, ATS-friendly and tailored to each posting: a fill-in outline, a full example, what screening software forgives and what it does not, edge cases and a FAQ.

• The French CV format: what to change when you apply for a job in France, with a full example

  Photo, one page, Bac+5, CEFR levels, the cover letter: what a US resume or UK CV must become to work in France, with a full annotated French CV example.

• The best free ATS resume checkers: 2026 comparison

  An honest comparison of free ATS resume checkers and free resume reviews: what each costs, whether you need an account, what the score measures, and where a person reviews your resume for free.