The phrase "beat the ATS" frames this as an adversarial problem, which is the wrong starting point. An Applicant Tracking System is not your opponent. It is a data-processing layer that sits between your application and a human, and it has very specific limitations. The goal is not to trick it. The goal is to ensure that your actual qualifications are legible to a system not designed with your career history in mind, on a platform that may be twenty years old, running a scoring algorithm configured by a recruiter in fifteen minutes.
Once you understand the mechanics, the path through it is straightforward. What follows is the current state of those mechanics, including the things that have changed significantly in the past two years and the advice that is still circulating from 2018 that will now actively damage your application.
What ATS Actually Does When It Receives Your Resume
Most job seekers think the ATS reads their resume the way a person does. It does not. It does two entirely separate things in sequence, and failing at either one produces the same outcome: your application never surfaces in a recruiter's queue.
Stage one: Parsing
When you submit a resume, the first thing the ATS does is convert your document into raw structured data. It strips out every visual element, every design choice, every formatting decision, and tries to extract discrete fields: your name, your contact information, your job titles, your employers, your employment dates, your educational qualifications, and your skills. It is doing this mechanically, reading left to right and top to bottom, categorising text into predefined boxes.
The problem is that this process is imperfect and sensitive to the way your document is constructed. Technical issues cause 43% of ATS rejections, with parsing errors accounting for 23% and formatting problems 12%, according to EDLIGO research published in 2025. When the parser encounters a multi-column layout, it reads both columns simultaneously from left to right, which means the text from your left column and your right column gets merged into a single garbled stream. When it encounters a table, it often cannot determine which cell belongs to which field. When it encounters a text box, a header, a footer, or a graphic, it either garbles the content or skips it entirely. None of this is visible to you. Your application status simply shows as received, and the submission that reached the recruiter is missing significant portions of your career history.
This is the stage where a large number of qualified candidates are eliminated before a single human decision is made, before keywords are even evaluated, before any assessment of fit occurs. The formatting of your document is doing more to determine your outcome than most people realise.
Stage two: Scoring and ranking
Once the ATS has successfully parsed your resume into structured data, it scores it. This is the stage most guides focus on, but they typically misrepresent how the scoring works in one critical way.
Resumes are scored on a 0 to 100 scale and ranked against all other applicants. Recruiters typically see a ranked list and start reviewing from the top. If your score falls below a threshold, often set by the recruiter, your resume is filtered out. The critical point is that the score is relative, not absolute. Your resume is not being evaluated against a fixed standard. It is being ranked against every other application received for that specific role, on that specific day, with the threshold set by that specific recruiter. A score of 72% might place you in the top 10 applicants for one role and in the bottom half for another.
Services that demand perfect scores are selling anxiety, not results. A well-tailored resume with 70 to 80% keyword alignment will consistently outperform an artificially inflated one. Understanding this changes the strategic logic entirely. You are not trying to hit an abstract benchmark. You are trying to score higher than the other qualified people applying to this specific role. That is a different and more tractable problem.
The Five Platforms That Screen Most Applications, and How They Differ
Not all ATS platforms are the same, and the advice that works well on one can produce a parsing failure on another. Workday, Greenhouse, Lever, iCIMS, and Taleo collectively process applications for roughly 78% of US Fortune 1000 hiring, according to HR Tech industry analysis from 2025. Each has distinct parsing logic, scoring behaviour, and formatting preferences. Knowing which one you are submitting to, which is often identifiable from the job portal URL or application flow, allows you to make targeted decisions rather than generic ones.
Workday is used by large enterprise companies across financial services, retail, and professional services. Its parser is strict on work history dates and requires the month and year format to be consistently applied throughout your document. Tables are a known failure point in Workday. Multi-column layouts cause parsing failures in Workday specifically. DOCX is the safer file format for Workday submissions, particularly in older configurations, because it does not carry the font-embedding risk that can cause PDF parsing failures. The AI screening layer in recent Workday configurations also evaluates skills alignment and career trajectory beyond simple keyword matching.
Greenhouse is common in growth-stage and mid-size tech companies and is generally the most resume-friendly of the major platforms. It parses text linearly and places heavy emphasis on the Skills section matching. PDF is generally safe in Greenhouse. The important caveat is that Greenhouse integrates with third-party AI scoring tools, which means the actual ranking logic varies by employer configuration. Two companies, both using Greenhouse, may be scoring candidates against entirely different criteria depending on what scoring modules they have enabled.
Taleo, owned by Oracle, is the oldest widely deployed system and the strictest in its parsing requirements. It is still in use at many large enterprises that have not migrated to more modern platforms. Tables, text boxes, page headers and footers, and embedded graphics still confuse Taleo's parser in 2026. DOCX is strongly preferred over PDF in Taleo environments. If the company you are applying to is a large traditional enterprise, particularly in manufacturing, hospitality, or consumer goods, there is a meaningful probability that it is using Taleo, and your formatting choices should reflect that.
Lever is popular with mid-size technology and professional services firms. Lever uses tag-based and full-text recruiter search rather than weighted keyword scoring. This means the recruiter searches for candidates using terms, and the system surfaces results where those terms appear close together in the source text. The practical implication is that a dense, clearly structured skills section at the top of your document, with skills listed in proximity to each other, performs better in Lever than in other platforms. PDF and DOCX perform comparably in Lever.
iCIMS is common in retail, healthcare, and manufacturing. It is an older platform with stricter parsing requirements than Greenhouse or Lever. DOCX files parse more reliably than PDFs in iCIMS environments, and table-based resume layouts frequently break.
How to identify which ATS a company uses before you apply
The job portal URL is often the clearest signal. Application portals hosted on workday.com, greenhouse.io, lever.co, icims.com, or taleo.net identify themselves directly. If the URL does not make it obvious, the application interface itself often carries visual branding from the ATS provider. Glassdoor reviews sometimes mention the system by name, particularly in reviews that discuss the application process. For companies where you cannot identify the system, a single-column DOCX formatted with standard headings and no tables or graphics will parse correctly in every major platform and is the safest universal choice.
The Formatting Failures That Kill Your Application Before Keywords Are Read
The multi-column problem
Multi-column layouts are the single most common parsing failure point across Greenhouse, Lever, and Workday. The system reads left to right and top to bottom, so a two-column layout causes it to mix content from both columns into a single garbled stream.
A resume that looks clean and professional in a design tool or PDF viewer can produce entirely incoherent output when parsed by an ATS. Your job title from the left column and your contact information from the right column get interleaved. Your skills section gets merged with your employment dates. The recruiter, if the application surfaces at all, sees a document that is difficult to read and appears to be missing sections. The fix is a single-column layout throughout, with the sole exception of a narrow contact information header at the very top of the document, which all major parsers handle correctly.
Tables, text boxes, headers, footers, and graphics
Each of these elements creates a specific failure. Tables are either skipped entirely or have their content scrambled because the parser cannot determine the relationship between cells. Text boxes are treated as images by most parsers, and the text inside them is lost. Content placed in document headers and footers, which is a common location for contact information in some templates, is extracted separately from the body content or not at all, meaning the recruiter may never see your email address or phone number. Graphics, including logos and icons, either produce error characters or are silently omitted. The practical rule is that if you cannot select the text in your resume by dragging a cursor over it in a plain text environment, the ATS cannot read it.
The plain-text paste test
This is the most reliable self-audit available, and it takes under two minutes. Copy the entire content of your resume and paste it into a plain text editor, Notepad on Windows or TextEdit in plain text mode on Mac. What you see is approximately what the ATS parser will extract from your document.
If your sections appear in the wrong order, you have a multi-column or table layout issue. If content is missing entirely, it was placed in a text box, a header, a footer, or a graphic. If bullet points have become random symbols or characters, your font choices or special characters are causing encoding problems. Any of these failures means your application is being submitted with corrupted or incomplete data, and no amount of keyword optimisation will compensate for that.
File format: the definitive answer
Modern ATS platforms handle both PDF and DOCX with high accuracy when the PDF contains embedded text, uses standard fonts, and has no copy protection. Independent testing published in 2026 shows a 96.7% parsability rate for standard text-based PDFs. The risk is in PDF type: a live text PDF exported from Word or Google Docs parses well. A PDF exported from Canva with custom fonts that are not embedded, or a scanned resume saved as a PDF image, can fail.
The practical recommendation is to use DOCX as the default when the job posting does not specify a format. It is universally safe across old and new ATS configurations. Use PDF when the employer specifically requests it or when you know the company is using Greenhouse or Lever, where both formats perform equivalently. Never submit a PDF created by a design tool without first verifying that the text is selectable in the file.
Keyword Strategy in 2026: What Has Changed and What Has Not
Why exact matching still matters even though ATS now uses NLP
Modern systems, including Workday's AI screening layer, Greenhouse's candidate scoring, and iCIMS's Talent Cloud, use natural language processing models trained on millions of job descriptions and resumes. They understand that "Python programming" and "Python development" refer to the same competency. This is genuinely useful. You do not need to use the exact phrase from a job description in every instance, and synonyms within the system's training data will register as matches.
However, exact matches still carry more weight in the scoring algorithm than inferred matches, particularly in older systems like Taleo that retain legacy exact-matching logic alongside newer semantic layers. The safe approach is to include both the exact phrase from the job description and your natural vocabulary when the two differ, woven into your bullet points rather than listed redundantly.
Where to place keywords for maximum scoring weight
Keywords must be relevant and naturally included to pass ATS and appeal to recruiters. The hierarchy matters because ATS scoring algorithms weight keyword appearances differently depending on where in the document they appear. A keyword in your job title carries more weight than the same keyword in your third bullet point. Your professional summary is the second-highest-weighted section. The skills section is the third. Keywords buried deep in the body of the document contribute to your score but less significantly than those appearing early and in structured sections. Aim for 15 to 25 relevant keywords with 60 to 80% coverage of job description keywords placed in high-priority zones: your job title, your professional summary, your skills section, and the first bullet point of each role.
What keyword stuffing does to your score in 2026
This is where a significant portion of job seekers are operating on advice that is now five years out of date. Resumes with over 20 skills listed separately suffered a 67% rejection rate, compared to a 34% rejection rate when those skills were integrated logically within work experience sections, according to EDLIGO research. Modern ATS detects and penalises keyword stuffing. Scoring algorithms have shifted emphasis from simple keyword counting to contextual relevance.
The mechanism is that modern ATS uses AI to detect unnatural language patterns. Repeating the same phrase multiple times, listing skills without contextual evidence, or using the same keyword in different grammatical forms in proximity all trigger low-intent flags. The system interprets these patterns as manipulation attempts rather than genuine descriptions of competency. Hidden text tricks, placing keywords in white text or in invisible sections, were obsolete by 2020. Modern ATS and AI review tools detect hidden text, and using it will get an application flagged for fraud and permanently disqualified.
Closing the vocabulary gap without misrepresenting your experience
The most common keyword problem among experienced professionals is not dishonesty but vocabulary mismatch. The terms used inside a company to describe a role are often different from the industry-standard terms used in job descriptions. A professional who spent three years doing what is described internally as "partner relationship management" may be applying for roles that use "strategic alliance management" or "channel partnerships" to describe the same function. The ATS scores these as different competencies.
Keywords are a translation layer between your experience and what the ATS algorithm is programmed to recognise. The fix is to identify the vocabulary the market uses for the function you performed, confirm that your experience genuinely corresponds to it, and describe your work using that vocabulary. This is not misrepresentation. It is translation, and it is the most high-leverage edit most experienced professionals can make to an existing CV.
The Professional Summary and Skills Section in 2026
Why the skills section is now the first section the ATS maps
According to LinkedIn's 2025 Future of Recruiting report, more than 60% of US enterprise hiring teams now filter candidates by specific required skills before reviewing job history. On Greenhouse and Workday, the Skills section is the first section that the parser maps to scorecard criteria.
This is a structural shift that most CV advice has not yet caught up with. For years, the professional summary was treated as the most important section because it was what a human recruiter saw first. In the ATS stage, the skills section is now the primary matching surface. The practical consequence is that a skills section which is sparse, generic, or absent does significant damage to your score before the recruiter ever reads your summary. The skills section should contain between 10 and 15 skills, described in the vocabulary of your target roles, and presented without tables or columns. Each skill listed should also appear in context within your work experience bullets, because an isolated skills list without contextual evidence is increasingly flagged by modern semantic matching systems as unverified.
How to write a summary that passes the machine and compels the human
The professional summary has two jobs that do not entirely overlap. For the ATS, it is the second-highest-weighted keyword zone in the document and should contain your most important role-specific terms, your seniority level clearly stated, and the sector or function you are targeting. For the human recruiter, it has approximately six seconds to communicate who you are and why this role makes sense, which means it needs to be specific, evidence-grounded, and completely free of generic phrases that signal a candidate who applies to everything.
The overlap between these two requirements is more significant than the tension. A summary that opens with your actual job title, names the sector you work in, quantifies one significant outcome from your career, and states the type of role you are targeting will satisfy both the ATS keyword requirement and the human readability requirement. What fails both simultaneously is the generic opener: "results-driven professional with extensive experience seeking a challenging role." This contains no scoreable keywords and communicates nothing to a human reader.
ATS for Senior and Executive Candidates: Why the Problem Is Harder, Not Easier
Why senior professionals assume ATS does not apply to them (and why they are wrong)
A common assumption among candidates at the Director level and above is that ATS is a problem for junior applicants and that senior hiring happens through networks and referrals. There is partial truth in this. Referrals do bypass some of the ATS friction. Even referred candidates in enterprise hiring pipelines are typically entered into the ATS by the recruiting coordinator, and their profile is scored against the role requirements before the hiring manager reviews it. The referral changes the probability of human review. It does not eliminate the ATS stage.
Executive and VP-level ATS filters are specifically tuned to find deep conceptual expertise through complex NLP, making semantic context more vital, not less. The system is looking for evidence of authority, not just keyword presence. The practical implication is that senior candidates face a different and, in some ways, harder ATS problem. The system is not just counting keywords. It is looking for patterns of language that demonstrate sustained high-level engagement with the competencies the role requires. A CV that describes fifteen years of work at a strategic level but in the vocabulary of internal documentation rather than market-standard leadership language can score poorly even on a modern semantic system, because the evidence of seniority is not expressed in the terms the system was trained to associate with senior-level competency.
How complex career histories interact with skills taxonomy systems
Modern ATS platforms use structured skills taxonomy frameworks, with systems like EMSI Burning Glass and O*NET used to classify skills from resumes. Rather than matching raw text, the system maps your skills to taxonomy nodes and checks for overlap with the job's required skill nodes. Skills that are not in the taxonomy may not register at all, even if they are clearly relevant. This is particularly true for emerging tools, domain-specific software, and non-English skill names.
For professionals whose careers have crossed multiple sectors, geographies, or disciplines, this creates a specific problem. The taxonomy was trained primarily on US and UK English job descriptions in conventional industry categories. A career built across pan-African programme management, Gulf fintech operations, and UK ecosystem building contains genuine and transferable skills that the taxonomy either maps incorrectly or fails to map at all, not because the skills are absent but because the vocabulary used to describe them falls outside the system's training data. The response is deliberate translation: identifying what market-standard term the taxonomy would use for a competency you have demonstrated and incorporating that term accurately into your document.
The global professional's specific ATS problem
Candidates with careers built outside the US and UK face a specific ATS disadvantage that is not addressed anywhere in mainstream career advice. ATS skills taxonomies are overwhelmingly trained on English-language job descriptions from US and UK markets. Non-Western certification naming conventions, company names outside the taxonomy's training data, and skills described in the professional vocabulary of a different market can fail to register as matches even when the underlying competency is exactly what the role requires.
The practical response is twofold. First, include both the local credential name and its closest internationally recognised equivalent where one exists. Second, describe the scope and impact of your work in vocabulary that the market you are applying to uses, not the vocabulary of the market where the work was done. This is not pretending the work happened somewhere it did not. It is presenting it in terms that the system and the recruiter can evaluate correctly.
AI-Generated Content and ATS Detection in 2026
How ATS now flags machine-generated language
The mass adoption of AI writing tools from 2023 onwards produced a specific market problem. Auto-apply platforms began generating cover letters and resume bullet points at scale, and the content they produced was characterised by generic phrasing, vague impact statements, and the kind of polished but empty language that sounds impressive without containing specific evidence.
ATS systems in 2026 include red flag detection for AI-generated content. According to LinkedIn's 2025 Global Talent Trends report, over 90% of recruiters now use AI-powered screening tools, and the shift in what those tools are looking for has moved firmly away from keyword density and toward evidence of competence. Recruiters are encountering AI-generated applications at such a volume that the generic language patterns have become reliably identifiable. The detection mechanism works on the same semantic analysis that the ATS uses for skill evaluation. Generic phrases like "spearheaded cross-functional initiatives to drive organisational transformation" contain no verifiable evidence of competency, no specific metric, no named outcome, and no contextual detail. Modern systems flag this pattern as low-intent content, which reduces your relevance score regardless of how well the surface keywords match.
How to use AI tools to assist without triggering detection
The distinction is between AI as a drafting assistant and AI as a submission tool. Using an AI tool to help structure a bullet point or to suggest the market-standard vocabulary for a competency you have genuinely demonstrated is legitimate and does not produce flaggable content, because the specific evidence, the numbers, the outcomes, and the context come from your actual career history and cannot be replicated generically. Using an AI tool to generate bullet points wholesale, without grounding them in specific evidence, produces the generic language patterns that modern ATS detection is specifically trained to identify.
The practical test is straightforward: if a bullet point on your CV could appear on the CV of a thousand other people in your field without any of them being dishonest, it is not doing its job for you, and a modern semantic ATS will score it accordingly.
The Self-Audit Before You Apply: Three Tests, Twenty Minutes
Before any application is submitted, three tests applied to your existing CV will identify the most significant problems and allow you to fix them without guesswork.
The keyword gap test requires you to take the full text of the job description you are targeting and compare it directly against your CV. Highlight every skill, qualification, technical term, and functional responsibility mentioned in the posting. Then check whether each term appears in your document, where it appears, and whether it appears in context within your work history or only in an isolated skills list. The gap between what the job description requires and what your CV currently demonstrates is your scoring deficit. Closing it with accurate, evidence-grounded language is the highest-leverage edit available before submission.
The platform identifier test uses the job portal URL, the application interface branding, or Glassdoor company reviews to identify which ATS the employer is using. Once identified, apply the platform-specific guidance from the section above to your formatting and file format choices. If the platform cannot be identified, default to a single-column DOCX with standard headings throughout.
The plain-text test, described in the formatting section above, remains the most important of the three for candidates whose CVs were built using design tools, two-column templates, or online resume builders that optimise for visual appearance rather than parse reliability. Run it every time you update your document, because formatting changes introduced during editing can silently reintroduce the problems you fixed.
How Candoorai Approaches This Problem
The three tests above describe the manual process for understanding how your CV will perform in an ATS before you submit it. What Candoorai does is run a version of that analysis automatically, at the role-specific level, from the same data source that then powers everything else in your search.
When you upload your CV and the URL of a target role, the platform analyses both simultaneously and tells you exactly how your experience maps to that specific role on that specific platform, what the ATS will score before a recruiter reads your name, and what the recruiter will see in the first six seconds once your document does reach a human. The output is not a generic resume score. It is a role-specific assessment that identifies the precise gap between how you are currently presenting your experience and how you need to present it for this application to score competitively.
For experienced professionals with complex career histories, this is the part of the job search that has historically required either expensive career coaching or an insider connection to someone who understands how the hiring system works. Understanding how your specific career reads to the system, filtering it, before the application goes anywhere, is the difference between searching with intelligence and searching blind. The same analysis then powers your interview preparation and your referral mapping, so that everything in your search is built from the same accurate picture of how your experience aligns with the role you are targeting.
Start free at candoorai
Understanding ATS is not a technical skill reserved for recruiters. It is practical knowledge that determines whether your qualifications reach a human being. The system has specific limitations. It was not designed with complex, global, or non-linear career histories in mind. Knowing where it fails and ensuring your document does not trigger those failures is the most direct thing you can do to improve the ratio between the quality of your experience and the number of interviews it produces.