You spent three hours on your resume. You're proud of it. You submitted it to 20 companies. You heard back from zero. The problem almost certainly wasn't your experience — it was an algorithm that rejected you before any human saw your name.
Welcome to the world of Applicant Tracking Systems, or ATS. Understanding how they work is the single most important thing you can do to improve your job search results.
What Is an ATS?
An Applicant Tracking System is software that companies use to manage job applications at scale. When you submit your resume online, it almost never goes directly to a recruiter's inbox. Instead, it's ingested by an ATS — parsed, scored, and ranked against other candidates, often before a human sees it.
Major ATS platforms include Workday, Greenhouse, Lever, iCIMS, and Taleo. They're used by virtually every company with more than 50 employees. When you apply through a company's careers page, a LinkedIn "Easy Apply," or Indeed, you're almost always going through an ATS.
How ATS Scoring Works
ATS systems score resumes primarily on keyword matching. They compare the text of your resume against the text of the job description and calculate a match score. The higher your score, the more likely you are to be surfaced to a recruiter.
- Exact keyword matches score highest — if the JD says "React.js" and your resume says "React", that may not match
- Section parsing — ATS looks for standard sections: Experience, Education, Skills. Unconventional headings like "What I've Built" may confuse the parser
- File format — DOCX and clean PDFs parse best; heavily designed PDFs with columns and graphics often parse poorly
- Date formats — inconsistent date formats can break experience parsing
The 75% Rule
Studies consistently show that approximately 75% of resumes submitted to large companies are rejected by ATS before a human reviewer sees them. This isn't because most candidates are unqualified — it's because their resumes weren't optimised for the system reading them first.
The implication is stark: for every 100 people who apply to a role, 75 are eliminated before a single human makes a judgment about their fit. The resume that gets you the interview isn't necessarily the best representation of your skills — it's the one that best speaks the language of the job description.
"An ATS doesn't care how impressive your career is. It cares whether your resume contains the words it's been told to look for."
What Recruiters Actually See
When a recruiter opens their ATS dashboard, they see a ranked list of candidates sorted by match score. They typically start at the top and work down. Candidates below a certain threshold may never be reviewed at all, especially for high-volume roles.
The recruiter doesn't know you were filtered out. They just see a list of "qualified" candidates — the ones whose resumes the ATS decided were relevant. If you weren't on that list, you didn't exist.
How AI Resume Tailoring Beats ATS
The solution to ATS filtering is resume tailoring — specifically, making sure the language in your resume mirrors the language in the job description. This used to mean spending 30–45 minutes manually editing your resume for every application. AI eliminates that time cost entirely.
Resume-MCP reads the job description, extracts the highest-value keywords — technical skills, frameworks, methodologies, soft skills — and rewrites your resume bullets to incorporate them naturally. The result is a resume that reads well to humans and scores well with ATS, without stuffing keywords unnaturally.
The difference in callback rate between a generic resume and a properly tailored one can be the difference between 0 responses and 5 interview invites from the same 20 applications. ATS is the gatekeeper. AI is the key.
What's Coming: Live ATS Score Preview
An upcoming Resume-MCP feature will show a live ATS match score next to your tailored resume preview — broken down by section (skills, experience, summary) and per-keyword. If the JD mentions "Kubernetes" 4 times and your tailored resume only covers it once, you'll see the gap before sending the application. The score updates in real time as the AI rewrites.
Pair that with the upcoming multi-resume A/B testing feature — apply with two different tailored versions to roughly half of similar roles each, then see which version gets a higher callback rate based on Gmail reply tracking. The job search becomes a real optimisation loop with data, not a black box of hope.