AI Resume Screening: What Employers Actually Check | Talstak
Candidates7 min read
AI Resume Screening: What Employers Actually Check, and How to Prepare Your CV
Most CVs are not rejected by a robot. They are filtered by criteria an employer set, then read by a person. Here is what that process checks and how to prepare a CV that holds up at every stage.
When you apply through a company's careers page, your CV usually lands in an applicant tracking system (ATS) before anyone reads it. Some of those systems now include AI features that summarise, match or rank applications. In SHRM's 2025 Talent Trends research, 43% of the HR professionals surveyed said their organisation used AI for HR tasks, up from 26% in 2024, and recruiting was the most common area of use.
That does not mean a machine decides whether you get an interview. In most hiring processes, software organises and filters applications according to rules the employer sets, and a recruiter or hiring manager makes the call. Knowing which parts are automated, and which are not, tells you where your effort pays off.
What happens to your application after you click submit
Every employer configures its tools differently, so treat this as the typical sequence rather than a description of any specific product.
1. The CV is parsed
The ATS extracts text from your file and maps it to fields: name, contact details, job titles, employers, dates, education and skills. Parsing works best on plain, single-column documents. Text inside images, text boxes, headers and footers, tables and multi-column layouts is where extraction most often goes wrong. When a field is misread, a recruiter searching the database may not find you, and the profile they open can look incomplete.
2. Screening questions are applied
Many application forms ask direct questions: Do you have the right to work in this country? Do you hold this certification? Can you work on site three days a week? Employers can configure some of these as knockout questions, which remove applications that give the wrong answer. This is the most common form of genuinely automatic rejection, and it is based on your answers, not on the wording of your CV.
3. Recruiters search, filter and sort
Recruiters often search the applicant pool by job title, skill, location or language. If your CV says "people lead" but the recruiter filters for "engineering manager", a literal search may not surface you. Modern search tools are more forgiving than older ones, but you cannot rely on every employer's setup interpreting synonyms well.
Some platforms compare applications with the job description and produce a match score, a summary or a suggested shortlist. How these features weigh evidence is generally not public, and it varies by vendor and configuration. What they have in common is that they work from the text they can extract, so clear structure and specific, relevant evidence help here too.
5. A person reads the CV
A recruiter typically reviews the shortlist, often quickly, and then shares candidates with the hiring manager. This is where most decisions are actually made, and where specific achievements, scope and relevance to the role matter most.
Five myths about automated resume rejection
"75% of CVs are rejected by an ATS before a human sees them." This figure has circulated for over a decade, but it traces back to marketing by a resume-optimisation vendor, and no published methodology supports it. Use it as a warning sign: advice built on it is usually selling something.
"The ATS rejects you if you miss a keyword." Outside of knockout questions, rejection is usually a human decision, sometimes made from a filtered or ranked list. The real problem is rigid criteria. In the Harvard Business School and Accenture Hidden Workers study, 88% of employers surveyed agreed that qualified high-skills candidates are screened out because they do not match the exact criteria in the job description. The fix is relevance and clarity, not more keywords.
"Hidden keywords in white text beat the system." Parsers extract all text, so hidden text typically shows up in the plain-text version a recruiter can view. At best it does nothing; at worst it ends your application on credibility grounds.
"Never send a PDF." Text-based PDFs exported from a word processor are generally parsed without trouble. Scanned documents, image-based exports and heavily designed templates are the risk, whatever the file format. If an employer asks for a specific format, use it.
"AI detectors will flag your CV." There is no good public evidence that AI-writing detection is a standard ATS screening step. The real risk of a CV generated wholesale by AI is different: it tends to be generic, and it can contain claims you cannot explain in an interview.
Your rights when automation is used in the EU
If you apply to employers in the EU, data protection law applies to your application. Under Article 22 of the GDPR, you generally have the right not to be subject to a decision based solely on automated processing that significantly affects you, and where such decisions are allowed you can ask for human intervention and contest the outcome. Articles 13 to 15 also give you the right to be told whether automated decision-making is used and to request a copy of your data.
The EU AI Act classifies AI systems used to filter applications or evaluate candidates as high-risk. After the 2026 Digital Omnibus amendment, the main obligations for these systems apply from 2 December 2027. Employers that use such tools will then face requirements around human oversight, transparency and record-keeping. Until then, the GDPR is the rule that protects you most directly.
What recruiters look for once a person reads your CV
Software decides little; people decide a lot. When recruiters read a CV, four questions come up again and again:
Relevance: Does this person's recent experience match what the role actually needs, not just its title?
Skills in context: Are the required skills named, and is it clear where and how they were used?
Evidence: Are there outcomes, scope and scale rather than a list of duties? Numbers help when you can stand behind them.
Trajectory: Do the moves make sense, and is there a reason this role is a logical next step?
A CV that answers those four questions quickly does well with a human reader and gives any matching tool clear text to work with.
Before and after: three rewrites
These examples are illustrative. Use your own real figures, and only include numbers you can explain in an interview.
1. From duties to evidence
Before: Responsible for managing the supplier quality process.
After: Led supplier quality for 14 component suppliers; introduced an incoming-inspection plan that reduced line stoppages from supplier defects over two quarters.
The second version shows scope, ownership and an outcome a recruiter can ask about.
After: Built a Python test-automation suite for embedded firmware and integrated it into the team's CI/CD pipeline (GitLab CI), cutting manual regression testing before each release.
The keywords are still there, once each, attached to real work.
3. From a generic profile to a targeted summary
Before: Results-driven professional with strong communication skills, seeking new challenges.
After: Mechanical design engineer with six years in precision tooling for automotive suppliers; experienced in Siemens NX, tolerance analysis and supplier handover. Looking for a senior design role in high-precision manufacturing.
The summary now tells a recruiter, in two lines, which roles you are relevant for.
How to tailor your CV without stuffing it
Mirror the employer's terms for skills you genuinely have. If the posting says "stakeholder management" and you describe the same work as "client coordination", use their phrase where it is accurate.
Spell out acronyms once. Write "Failure Mode and Effects Analysis (FMEA)" the first time, so both literal searches and human readers recognise it.
Align job titles honestly. If your internal title was unusual, add the common equivalent in brackets, for example "Delivery Owner (Engineering Project Manager)".
Reorder, don't rewrite your history. Keep a master CV with everything, and for each serious application move the most relevant experience and skills to the top.
Answer screening questions carefully. They often carry more weight than anything in the CV itself.
Using AI tools on your CV responsibly
AI writing tools are useful as editors: tightening long bullet points, checking consistency, or suggesting how a skill could be phrased. They are risky as authors. If a tool invents an achievement or inflates your role, you are the one who has to defend it in the interview. Write the substance yourself, use AI to polish, and read every line before you send it.
Checklist before you apply
Single-column layout, standard section headings (Experience, Education, Skills), no text in images or text boxes.
Contact details in the body of the document, not in the header or footer.
A two- or three-line summary that names your field, seniority and the kind of role you want.
For each recent role: scope, your responsibility and at least one concrete outcome.
Required skills from the posting that you actually have, named explicitly and shown in context.
Acronyms spelled out once; job titles clear to an outsider.
Dates consistent, gaps explained briefly if relevant.
Screening questions answered accurately.
No hidden text, no claims you cannot explain.
A file the employer accepts, ideally a text-based PDF or Word document.
