What an ATS Actually Reads From Your Resume (2026)

See what the robot reads.

Check your resume free - No signup, no card. See what an ATS pulls from your CV, and score it.

What an ATS actually does with your resume

Most advice treats the applicant tracking system like a bouncer that says yes or no. It is closer to a librarian that files you, badly, and fast. When you upload a CV, the parser does not look at your layout. It pulls out a stream of text, tries to sort that text into fields, and stores the result. A recruiter then searches and sorts that database. Your real resume is sitting in a folder no one opens; the parsed version is what gets you found.

That means you have two gates, not one. First the parser has to extract your experience correctly. Then a human has to like what surfaces, in seconds, not minutes, on the first pass. Most candidates obsess over the second gate and lose at the first, because they never saw what the machine actually read.

What the parser keeps, and what it throws away

A parser wants plain, ordered text. It keeps:

• Real, selectable text in a single reading order
• Standard section headings it recognises (Experience, Education, Skills)
• Dates and job titles it can map to fields

It quietly loses:

• Text trapped inside images, logos, or icons (a parser reads zero characters from a picture)
• Anything in headers and footers on some systems
• Content in tables, text boxes, or a side column that breaks the top-to-bottom reading order

Here is the 5-second test you can run right now: open your PDF and try to select the text with your cursor. If you can highlight your name and copy it, the file has a real text layer and an ATS can read it. If the text will not select, it is an image, and the machine sees an empty page. Resumes exported from design tools like Canva or "printed to PDF" as an image fail this test constantly.

One column or two? The layout that decides everything

This is the single biggest formatting mistake, and it is invisible until a parser jumbles you.

| Layout | How an ATS reads it | Verdict |

| --- | --- | --- |

| Single column | Top to bottom, in the exact order you wrote it | Safe everywhere |

| Two column / sidebar | Often reads one column, then the other, splicing your skills into the middle of a job | Risky on strict systems |

A two-column resume looks sharp to a human reviewer. But a parser reading left-to-right, top-to-bottom can read your sidebar (skills, contact, education) separately from your main column, so your carefully ordered experience arrives scrambled. Modern parsers handle simple two-column layouts better than they used to, but "better" is not "reliably," and you rarely know which system a company runs. If a role screens heavily on software, a clean single column is the safe call. Save the two-column design for when you know a person opens the file directly.

PDF or Word? The honest answer

Both can work. The honest ranking, based on how the big systems behave in 2026:

• A text-based PDF (exported straight from Word, Google Docs, or a proper builder) parses well on modern systems like Greenhouse and Lever. Older enterprise systems such as Taleo and iCIMS are less forgiving and occasionally misread PDFs.
• DOCX (Word) is the most universally safe format, because its structure is tagged text that almost every parser reads in order. If a job board says "PDF or Word" and you are not sure which system sits behind it, Word is the low-risk default.
• An image or scanned PDF is the worst case. The parser extracts nothing.

The takeaway is not "never use PDF." It is: if you send a PDF, make sure it is a real text-layer PDF with a single-column, standard layout, then it behaves like Word on the systems that matter. When in doubt, or when the employer asks, send the Word version.

What silently breaks parsing (and you never get told)

No ATS emails you to say "we couldn't read your skills section." It just ranks you lower. The usual culprits:

• Two columns or a sidebar holding key information
• Tables used to lay out experience or skills
• Graphics, charts, logos, or icons standing in for text
• Creative section names ("Where I've Made Impact" instead of "Experience")
• Headers and footers holding your phone number or email
• Uncommon fonts that render as gibberish on extraction

None of these show up when you look at your beautiful resume. They only show up in the parsed text, which is exactly why you should look at the parsed text.

How to see what an ATS reads

You do not have to guess. Two checks take under a minute. First, the select-text test: open your PDF, try to highlight your name and bullets. If the text will not select, an ATS sees an empty page, fix that before anything else. Second, paste your resume into ApplyArc's free ATS checker. It shows you what a parser pulls out, the sections it detected, the formatting issues that trip it up, and a score, in seconds, with no signup. That is the gap between "I think it looks good" and "I know it reads clean."

Read your resume the way the software does.

Check your resume free - See what an ATS pulls from your CV, then score it against a real job. No signup.

The 60-second ATS-safe checklist

Run this before you hit submit:

Get those seven right and the machine reads you exactly as you intended. After that, the only thing left to win is the keyword match for each specific role, which is a different fight, and a winnable one. (And no, despite the scare stories, an ATS almost never rejects you outright, it ranks you.)