What happens to your CV after you press Apply?

It is not a robot. It is software, and mostly not even AI.

Almost everything written about applicant tracking systems starts with the words "a robot reads your CV". There is no robot. An applicant tracking system, or ATS, is a piece of software: a database of applicants with a workflow attached, the same kind of program a company uses to keep track of its customers or its invoices. It stores every application for every open job, records who is at which stage, and gives recruiters a search box.

Some systems now include artificial intelligence, usually to match applicants against a job and sort them into groups. That is still the minority. In January 2024, the Society for Human Resource Management (SHRM) surveyed 2,366 HR professionals in the United States. Only 26 per cent of their organisations used AI for any HR activity at all. Of those, 64 per cent used it somewhere in recruiting, interviewing or hiring, and about one in three of that group used it to review or screen CVs. Put the figures together and AI was reading CVs at roughly one organisation in twenty. Everywhere else, the software was running rules and searches that a person had set up.

How does an ATS work?

An applicant tracking system receives your application, extracts the text of your CV into a structured record, checks your answers against rules the employer has set, and then lets recruiters search, filter and rank everyone who applied. It does not read your CV the way a person does, and in most systems it does not decide who gets hired. It decides, in effect, who gets seen.

The rest of this article follows that journey in order: what the recruiter sets up before you apply, and what the software does with your CV afterwards. The details differ between products, so I have taken the steps from the official documentation of three widely used systems, Greenhouse, Workday and iCIMS, and from research that does not sell CV software.

The recruiter's side: what happens before you apply

To understand what the software does with your CV, start with what the recruiter does before you ever see the advert.

1. The job is created in the system. The recruiter opens a requisition: title, location, the job description, and the application form you will later fill in.
2. Knock-out questions are added. These are the yes-or-no questions on the application form: do you hold a Class A driving licence, are you allowed to work in this country, can you work weekends. In Greenhouse, the employer can mark certain answers as an automatic rejection; the help centre's own example is a driving job that requires the licence (Greenhouse, Auto-reject (opens in a new tab)). Workday does the same through what it calls automatic stage routing: rules built from the application form, the questionnaire and any assessment result, which advance or decline a candidate without a person looking (Workday Administrator Guide (opens in a new tab)). iCIMS describes pre-screening assessments that return a score which "qualifies or disqualifies a candidate" (iCIMS (opens in a new tab)). This is the one place where the software really does reject people on its own, and it is triggered by your answers, not by your CV.
3. Must-have criteria are set. The recruiter can also tell the system which requirements to filter on: a bachelor's degree, a professional licence, a certification. Harvard Business School's 2021 study of 2,275 executives in the United States, the United Kingdom and Germany found that these filters often rely on proxies, such as a degree standing in for skills, and that almost half of the companies surveyed automatically screened out any CV with a gap in employment of more than six months (Hidden Workers: Untapped Talent (opens in a new tab)).
4. The advert goes out, and the applications start arriving in the same system.

Your side: what happens after you press Apply

1. Your CV is parsed. The system runs your file through a parser, a program that tries to turn the document into fields: name, contact details, each job with employer, title and dates, education, skills, languages, licences. Textkernel, whose parser is built into many of these systems, says in its own developer documentation that columnar layouts can confuse the parser and that some sections, such as patents and publications, are "impossible to parse at a granular level" (Textkernel, Parser Output (opens in a new tab)). Whatever the parser cannot place ends up in the wrong field, or nowhere.
2. Your answers are checked. If the job has knock-out questions and you gave a disqualifying answer, the application is rejected now, usually with an automatic email. If not, you move on.
3. You become a record in the database. From here on you are a row in a table: the parsed fields, your original file attached, a stage such as "New" or "Phone screen", and whatever notes and tags the recruiter adds.
4. The recruiter searches, filters and ranks. This is the step most people mean when they say "the ATS rejected me", and it is the one that is least automatic. Indeed's guide for job seekers describes it plainly: recruiters search the applications for keywords taken from the job description, such as particular skills, job titles or responsibilities, and the system can rank applicants by how closely their details match the requirements (Indeed Career Guide (opens in a new tab)). A CV that does not contain the words the recruiter searched for is not deleted. It simply does not come up. Systems with AI matching go one step further: Greenhouse's Talent Matching sorts applicants into Strong, Good, Partial, Limited or Needs manual review, measured against criteria the hiring team wrote, and Greenhouse's own FAQ states that it "does not automatically disposition any candidates nor does it make any hiring decisions" (Greenhouse, Talent Matching FAQ (opens in a new tab)).
5. A person looks, briefly. When a recruiter does open your CV, the first look is short. In 2018 the careers site Ladders fitted recruiters with eye-tracking equipment and measured an average of 7.4 seconds for the initial scan. The CVs that held attention had simple layouts with clear section headings, bold titles and bulleted achievements. Cluttered pages, multiple columns and long sentences lost it (HR Dive on the Ladders study (opens in a new tab)).

What it means for the recruiter

From the recruiter's chair, the software is mostly about time. In SHRM's 2024 survey, 88 per cent of the organisations using AI in recruiting gave saving time as their reason (SHRM, 2024 Talent Trends (opens in a new tab)). Two years earlier, in a SHRM survey of 1,688 HR professionals, 64 per cent said their automation or AI tools automatically filtered out unqualified applicants, and 19 per cent said the tools had excluded qualified ones (SHRM, April 2022 (opens in a new tab)).

The Harvard study puts it more bluntly. An ATS, it says, "largely relies on 'negative' logic to winnow the applicant pool": people are removed for what they lack, a degree, continuous employment, an exact phrase, rather than considered for what they have. Eighty-eight per cent of the executives surveyed agreed that qualified high-skills candidates are vetted out of the process because they do not match the exact criteria. The systems are doing what they were bought to do. They were not bought to find you.

What it means for you

None of this means you can beat the system, and this article is not going to tell you how. It means a CV now has two readers, in a fixed order: first a parser and a search box, then a person with 7.4 seconds. A CV that works has to satisfy both.

• It has to be readable by the parser. Every job needs to come out as a job, with the employer, the title and the dates attached to it, and every section needs to land in the right field. That is a question of structure, not of design.
• It has to contain the words the recruiter will search for. Those words are in the job advert. If the advert says "stakeholder management" and your CV says "worked with many departments", the search will not find you, however true both are.
• It has to answer the knock-out questions honestly. They are the one automatic rejection, and a licence or a work permit you do not have will surface later anyway.
• It has to explain its gaps in words. The filter looks at dates. The person who reads your CV afterwards reads the sentence next to them.

The first two points are the part that software can do for you, and it is the part Adapt to Success does: it keeps your facts and rebuilds the structure and the wording so that the parser and the search box can read them. The last two are yours.

Who is watching the watchers?

Two things are worth knowing if you apply for jobs in Europe or the United Kingdom.

The European Union's Artificial Intelligence Act lists AI systems used to recruit or select people, including systems that filter job applications and evaluate candidates, as high-risk in its Annex III (EU AI Act, Annex III (opens in a new tab)). That puts them under the Act's obligations for risk management, data quality, human oversight and transparency.

In November 2024, the United Kingdom's Information Commissioner's Office published the outcome of an audit of AI tools used in recruitment. It made almost 300 recommendations. Some of the tools had let recruiters filter candidates by protected characteristics, some inferred gender and ethnicity from names, and some collected far more personal data than the job required (ICO, November 2024 (opens in a new tab)).

The point is not that the software is illegal. The point is that it is now regulated, audited and, in the EU, classed as high-risk, which tells you how much weight it carries in whether you are seen.

What I wish I had known

I did not know any of this when I sent my 2,200 applications. I formatted my CV to look good to a person, and most of the time no person saw it. If you want to see what the difference looks like on an actual page, there are four before-and-after examples on the ATS Explained page. If you want to know how often this software stands between you and a recruiter, that is a later article in this series.