The short version
- Mostly no. Neither Workday nor Greenhouse documents a score cutoff that deletes a resume, and no employer publishes a threshold.
- The main documented automated rejection runs on the application questions. A recruiter-set rule rejects a disqualifying answer, on work authorisation for example, without review.
- Where employers use them, ranking tools like Workday's HiredScore and Greenhouse Talent Matching are documented to sort applicants rather than remove them. The order decides who gets read.
- An unreadable resume is misfiled, not rejected. The damage is empty fields that a recruiter's search cannot match.
- Beyond that, Greenhouse tells candidates that, by and large, applications in Greenhouse are reviewed by real people, while conceding that some companies use AI to auto-reject without human input.
It is the single most repeated belief in job searching: submit a resume, an algorithm scores it, and if the number is too low the file is discarded before a person sees it. Whole products are sold against that fear. It is worth checking against what the companies that build these systems actually say.
They describe several separate mechanisms, and the application questions are the main one documented as rejecting anyone automatically. Here is what each one does, and where we cover it in depth.
| Mechanism | Rejects automatically? | What it actually decides |
|---|---|---|
| Knockout questions | Yes, if a recruiter sets a rule | Whether you remain a candidate |
| Match ranking (Workday, Greenhouse) | No, it prioritises | Where you sit in the recruiter's list |
| Resume parsing | No, it fills fields for a person to review | Which fields your experience lands in |
| Recruiter keyword search | No, a person writes the query | Whether you appear in the results |
| Recruiter review | It is the decision, made by a person | Whether you advance |
Knockout questions: the main documented automated rejection
The main mechanism documented as rejecting applications without human review is a set of application rules attached to screening questions, usually called knockout questions.
A recruiter selects a question and a disqualifying answer, and matching applications are rejected automatically. Greenhouse's page for candidates gives the plainest example: answer "no" to whether you are legally allowed to work in the country, and the application is rejected. It fires on the form, not on the attachment.
In Greenhouse it works only on Yes/No, single-select and multi-select questions, and only on the Plus and Pro tiers. The hiring team is not notified when it fires and the rejection email is optional, which is how an application can return nothing at all. Our knockout questions guide covers the Workday equivalent and how to answer safely.
The distinction matters. A machine rejecting your resume on its content and a dropdown answer matching a rule a person wrote call for completely different responses, and Greenhouse's and Workday's product documentation describes only the second.
Ranking tools sort applicants rather than remove them
Match scoring is real. Employers that use Workday's HiredScore grade applicants with it, which Workday describes as prioritising candidates for the recruiter. Greenhouse Talent Matching, an add-on on newer plans, sorts applicants into categories such as Strong, Good, Partial, Limited or Needs manual review.
Neither is a gate. Greenhouse states that Talent Matching does not auto-reject or auto-advance any candidate, including those in lower categories, and that recruiters and hiring managers make every advancement and rejection decision. Nothing we found in Workday's documentation describes its grade removing anyone, and Workday does not publish the model behind it.
What a grade decides is reading order, which decides whether anyone opens your file. We cover how Workday's grading works and how Greenhouse and Lever rank applicants separately.
What happens to a resume the parser cannot read
A file the system cannot read is not thrown away either. Greenhouse's documentation tells the recruiter to enter the details by hand when a parse fails, and Workday describes parsing as filling fields for a person to review.
The cost is quieter: titles and dates that land in the wrong field, or in none, cannot match a recruiter's search. Our guide to how layout breaks ATS parsing covers the layouts that cause it, and an ATS resume checker flags the sections and contact details a parser could not find.
Keyword matching is mostly a recruiter searching
Much of what gets called keyword filtering is a recruiter typing a Boolean query. A person decides what matters, and you are either in the results or not. No vendor documents a keyword density score. We cover where to put terms so they count in our keyword matching guide.
What Greenhouse says happens after the form
Past the form, Greenhouse documents leaving rejection to people. Its page for candidates says applications in Greenhouse are, by and large, reviewed by real people, one by one, in the order received. On its AI, it is blunter still.
“AI doesn't score or rank applications, nor does it make any decisions about whether or not you move forward.”
The same page is candid about other employers: it concedes that "some companies use AI to analyze, score and auto-reject applications without any input from humans." Greenhouse's own automated filters, it says, are set and controlled directly by recruiters, described as living, breathing humans, rather than outsourced to a bot.
That page is dated January 2025, and the ranking line has aged: Greenhouse now documents Talent Matching, which does sort applicants. The decision line still holds for Greenhouse's AI: Talent Matching leaves every advancement and rejection decision to recruiters and hiring managers.
Why the myth survives anyway
Because the experience it describes is accurate even though the mechanism is not. You apply, nothing happens, the posting stays open for weeks. Something clearly went wrong and no one tells you what.
Ranking explains that experience. A recruiter reads down a sorted list until the shortlist is full, then stops. Nobody at the bottom was rejected. They were never reached, and from outside the silence looks the same.
“The myth gets the feeling right and the mechanism wrong, which is exactly why it is so hard to kill.”
It also survives because it sells. The most quoted figure, that 75% of resumes are rejected by an ATS, traces to a 2012 sales pitch by Preptel, a resume-optimisation vendor that shut down the following year. A score you can raise is easier to package than ranking against a specific requisition.
Check how your resume parses
Optimize your application materials in under 2 minutes.
To be clear about what a tool like this can and cannot tell you: it reports what a parser could and could not extract from your file. It cannot show you a score inside an employer's system, because that score is not exposed to candidates, and any product claiming otherwise is estimating.
What actually follows from this
Dropping the auto-rejection model changes where the effort goes.
- Stop chasing a match percentage. No employer publishes a threshold.
- Slow down on the application questions. They are the best-documented way to be rejected without review.
- Make the document extractable, because ranking and search both run over parsed text.
- Write for the person who opens the file. In Greenhouse's and Workday's documented flows, a person makes the decision.
- Treat file format as settled. It matters far less than it is said to.
You are not trying to defeat a filter. You are competing for position among other applicants, in front of a recruiter who will run out of time before they run out of applicants. That is the problem actually in front of you.