The short version
- An ATS is a database and a workflow tool for hiring teams. It parses your resume into text and fields so recruiters can search it.
- Most rejections are made by people using the ATS’s search and filters, not by an algorithm silently discarding you.
- The failures that are genuinely the software’s fault happen at parsing: columns, images, icon fonts and odd headings.
- The fix is boring and reliable: one column, real text, standard headings, and the posting’s own vocabulary.
What an ATS is
An applicant tracking system is the software a hiring team uses to run its hiring: posting jobs, receiving applications, moving candidates through stages, scheduling interviews and keeping records. From your side it is the careers page and the “Upload your resume” button. From the recruiter’s side it is a database of candidates with a search box.
Common examples include Workday, Oracle Taleo and iCIMS at large employers, and Greenhouse, Lever, Ashby and SmartRecruiters at many tech companies. They differ in details, but they all do the same two things with your resume: parse it, and make it searchable.
Step 1: parsing
When you upload a file, the ATS extracts its text and tries to split it into fields: your name and contact details, each job with its title, company and dates, your education, your skills. This is the step where software can genuinely fail you, because extraction is literal. It reads the text in the order the file stores it, which is not always the order your eye reads the page.
A two-column layout is the classic casualty: the extractor reads straight across the page, so a line of your skills sidebar ends up glued to a line of your job history. A contact block drawn as an image is simply absent. An email address next to an icon-font envelope comes out next to a junk character.
Step 2: search, filters and a person
Once parsed, your resume joins every other applicant’s in the system. A recruiter then does what anyone does with a database: they search it. They type the skills and titles that matter for the role, filter by location or years, sort, and open the profiles that look promising.
This is why the posting’s vocabulary matters. If the job asks for “stakeholder management” and your resume says “worked closely with teams across the business”, a search for the posting’s phrase will not find you — even if the work was identical. Some systems match synonyms; you should not rely on it.
Many systems also offer ranking or “match” features, which compare resumes against the job description. How each vendor’s ranking works is not public. Anyone who claims to know the exact formula Workday or Greenhouse uses is guessing.
The auto-rejection myth
You will see confident claims that most resumes are “rejected by robots” before a human sees them. There is no solid public evidence for the specific numbers usually quoted, and the picture they paint is misleading.
What does reject people automatically is usually the application form, not the resume: knockout questions about work authorisation, a required licence or certification, or a minimum number of years. Answer those honestly and carefully. After that, the more accurate picture is less dramatic and more useful: your resume is not thrown away, it is simply hard to find or hard to read.
What actually goes wrong
- Layouts that extract out of order — columns, tables used for layout, text boxes.
- Text that is not text — scanned resumes, designs exported as images, names set in a graphic.
- Unrecognisable headings — “My journey” instead of “Experience”, so jobs are not parsed as jobs.
- Icon fonts and decorative characters — they extract as junk next to your contact details.
- Letter-spaced headings — a heading styled as “E X P E R I E N C E” extracts as eleven separate letters, which nothing maps to a section.
- Dates a parser cannot read — creative formats, or roles that seem to end before they start.
- Missing vocabulary — the work is there, described in different words from the posting.
Every one of these is fixable, and our ATS-friendly format guide goes through them rule by rule.
How to check your resume
Do the copy-and-paste test above. Then run your resume through our free resume checker, which reads it the way our scorer does, and the job match checker against a posting you care about. Both run in your browser; nothing is uploaded.
In the AI-Resume app we go one step further: every PDF you export is read back with a text extractor and checked for reading order, headings, contact details, dates and glyphs before you send it. More on that on the PDF export page.
Questions
Which companies use an ATS?
Almost every company that hires at any scale. Large employers commonly use systems such as Workday, Oracle Taleo, SAP SuccessFactors and iCIMS; startups and mid-sized tech companies often use Greenhouse, Lever, Ashby or SmartRecruiters. If you applied through a careers page, an ATS almost certainly received your resume.
Does an ATS automatically reject resumes?
Usually it does not reject on its own. What does filter people out automatically are knockout questions in the application form — work authorisation, a required licence, willingness to relocate. Beyond that, recruiters search and sort; a resume that parses badly is simply harder to find.
Can an ATS read a PDF?
Yes, if the PDF contains real text. A PDF exported from a word processor or typeset by a resume tool is fine. A scan, or a design exported as one big image, has no text for the parser to read.