What Is an ATS? How to Write a Resume That Gets Through in 2026

An Applicant Tracking System (ATS) is the software companies use to receive, store, and search job applications. Almost every mid-size and large employer uses one. There's a lot of folklore about "beating the ATS" — most of it outdated. This guide explains what these systems really do in 2026 and how to write a resume that parses cleanly and reads well to the human who sees it next.

What an ATS actually does

At its core an ATS does three things: it parses your file into structured fields (name, contact, work history, skills), stores it in a database, and lets a recruiter search and filter that database — usually by keyword, title, location, and recency. Modern systems also rank or surface candidates, but a human still makes the call. The ATS is a filing cabinet with a search box, not a gatekeeper that auto-rejects you.

Understanding that shape explains most of the advice that follows. If the system is a search index, then your job is to be findable and legible — not to outwit an algorithm.

Does an ATS reject you automatically?

This is the single most common fear about these systems, and it is mostly misplaced. Most ATS platforms do not assign a secret match score that bins your resume below some threshold. There is no hidden number deciding your fate before a human is involved.

The real rejection path is duller and more fixable. A recruiter opens the system, searches for the two or three skills the role actually needs, and works through the results. If your resume does not contain those words — or the parser mangled the section they were in — you never appear in the list. Nobody rejected you. You were never retrieved.

There are narrow exceptions. Some employers set hard knockout questions in the application form itself: work authorization, a required license, minimum years. Those are answered by you in the form, not inferred from your resume, and they genuinely can filter you out. That is a form question, not the ATS reading your document.

The practical consequence: stop optimizing against an imaginary scoring robot and start optimizing for a keyword search run by a busy person. Those two goals produce very different resumes, and only one of them also reads well.

The systems you'll actually meet

"ATS" gets used as though it were one product. In practice you are dealing with a handful of platforms, and knowing which one you are in tells you what to expect from the application itself.

• Workday — common at large enterprises. Usually the longest applications, and often asks you to re-key your work history into form fields after uploading a file.
• Greenhouse, Lever and Ashby — widespread at technology companies and startups. Generally lighter application flows built around a file upload.
• Taleo (Oracle) and iCIMS — long-established enterprise systems, frequently behind the more dated application portals you'll meet at very large organizations.

The differences matter less than people assume. None of them is going to reject you on a hidden score, and all of them are ultimately search interfaces for a recruiter. Where they do differ is tolerance for unusual formatting — which is an argument for a conservative layout everywhere rather than a different resume per platform.

One thing worth doing regardless of platform: when an application makes you re-key your history into form fields, fill them in properly. Recruiters at those companies frequently search the structured fields, not the attached file.

Where resumes actually break

• Layout the parser can't read — multi-column layouts, text inside images, tables, and headers/footers often scramble or vanish during parsing.
• Missing keywords — the skills and tools in the job description aren't present in your resume, so keyword searches don't return you.
• Vague titles — a creative title like "Growth Ninja" is invisible to a recruiter searching "Marketing Manager".
• Wrong file type — a flattened PDF export or an unusual format the parser mishandles.

MIT's career advising team puts the formatting risk plainly, warning that these systems "may struggle with reading tables, columns, graphics, and images". That is the whole case for a conservative layout — and there is a fuller treatment of what a scannable resume looks like in how to format a resume recruiters can scan.

File formats: what to actually send

The PDF-versus-Word argument is older than it is useful. Both work in every mainstream system today. What matters is not the extension but whether the file contains real, selectable text.

• Text-based PDF — the safe default. Exported directly from your editor, with text you can select and copy. Preserves your layout exactly.
• .docx — perfectly fine, and the right choice when the posting explicitly asks for it. Slightly more prone to shifting layout on someone else's machine.
• Scanned or image-based PDF — the actual failure case. A photo or scan of a resume contains no text to extract, so a parser sees an empty document.
• Design-tool exports — check them. Some flatten text into vector outlines, which look identical to you and read as blank to a parser.

The one-second test: open your PDF, try to select a line of text and copy it. If you can paste it somewhere and read it, a parser can too. If nothing selects, you have an image and it will not parse.

How keyword matching really works

Recruiter search is far more literal than people expect. It is largely string matching — which is exactly why the advice to "mirror the job description" works, and why it is so easy to do badly.

Two rules cover most of it. Use the posting's exact term where it is genuinely true of you: if the job says Kubernetes, write Kubernetes, not just K8s. And when a term has a common abbreviation, use both at least once — "Kubernetes (K8s)" costs you four characters and matches either search.

The same applies to titles. A recruiter searching for "Product Manager" will not find "Product Owner & Growth Lead", however accurate that was internally. Lead with the standard title and put the internal one in parentheses if it matters to you.

This is also the honest way to read a coverage score. nxtleap's ATS audit computes keyword coverage by exact string matching in code rather than asking a language model whether your resume "seems relevant" — and labels every number in the audit with whether it was measured or inferred. A coverage figure you cannot trace back to specific words is not worth much. You can run that check on an existing resume with the ATS resume checker.

If you call the document a CV rather than a résumé, none of this changes — see CV vs resume for what actually differs. Two follow-ons go deeper than this section can: which keywords in a posting are actually load-bearing, and what a good ATS score means — including why the formatting half and the keyword half should never be averaged into one number.

An ATS-friendly resume checklist

1. Use a single-column layout with standard section headings (Experience, Education, Skills).
2. Submit a text-based PDF (or .docx if the posting asks) — never a scanned image or screenshot.
3. Mirror the exact terms from the job description where they're genuinely true of you — if it says "Kubernetes", don't only write "K8s".
4. Keep a real Skills section with the concrete tools and methods you know.
5. Use standard job titles; put any fun internal title in parentheses if you must.
6. Put contact details in the body of the document, not in the header or footer.
7. Fill in the form fields properly when an application asks you to re-key your history.
8. Put accomplishments in plain bullet points with real metrics — see how to quantify them.

Don't optimize yourself into a worse resume

The failure mode of "ATS optimization" is keyword-stuffing that passes a search but bores the human who reads it ten seconds later. The goal isn't to trick software — it's to be legibly relevant to both. Write for the person; format for the parser.

It is worth being blunt about one popular tactic: white text, hidden keyword blocks, and keyword lists buried in the margins. These are visible to anyone who selects all the text in your document, they are well known to recruiters, and they read as an attempt to deceive. The downside is not a lower score. It is being remembered as the candidate who tried it.

How nxtleap helps

nxtleap's score & tailor reads the job description, ranks your Achievement Bank against it, and assembles a clean, single-column, parser-safe resume focused on that role. Every template it ships is single-column by default and carries its own parseability rating, so the formatting decision is made before you start.

Two things run on the result. The ATS Audit reports keyword coverage by exact string matching and tags each field as measured or inferred — no invented confidence. And a gap analysis names the requirements in the posting your achievements genuinely do not cover, so you are told what is missing rather than having a plausible-sounding claim written in to paper over it. Each achievement also carries an Impact Score — Elite at 75 and above, Strong at 60–74, Task below — so you can see which bullets are earning their place before a recruiter has to.

Explore the ATS resume checker, see how the whole flow works, or read the companion guides on writing the bullets and formatting the page — or the complete guide if you are rebuilding from scratch. To check a resume against one specific posting, see resume vs job description.