Resume Keywords: How ATS Search Matches Them, Plus Role Lists

TL;DR: Resume keywords are the job titles, tools, certifications and methods a recruiter types into the applicant tracking system's search box. Some ATS searches match the exact words you wrote; newer AI matching also accepts close synonyms. Write for the strict case: use the job ad's exact wording for each skill you have, spell out acronyms once, and prove each term in a bullet. Then compare your resume with one real ad before you apply.

What are resume keywords, and where does an ATS use them?

Resume keyword: a word or short phrase a recruiter or matching tool looks for in your resume text, usually a job title, a hard skill or tool, a certification, a degree or a domain term named in the job ad.

An applicant tracking system handles your resume in separate steps, and keywords matter in only some of them. First the system parses your file into text and fields. Then a recruiter searches or filters the applicant pool, and in some products an AI feature sorts applicants against the job's criteria. Keywords do their work in that second stage.

Automatic rejection mostly happens elsewhere, through knockout questions. Greenhouse's own Project Manager posting, published on its Greenhouse job board on 9 October 2026, states that "applicants must be currently authorized to work in the United States on a full-time basis". A question like that can end an application before anyone reads the resume, and no keyword changes the answer.

The keyword stage is softer and more common. Greenhouse lets recruiters search resumes for keywords and shows a snippet of how the terms are used in each matching profile. Its Talent Filtering feature suggests keywords from the public job post and lets the recruiter mark each one "Preferred" (an OR) or "Required" (an AND). A resume that lacks a required term does not get deleted. It drops out of that recruiter's results, which for that search has much the same effect.

Getting into the text in the first place is a formatting question, covered in the ATS-friendly resume format guide. This post assumes the parser can read your file and deals with what it then searches for.

How ATS keyword matching works: exact phrases, wildcards and semantic match

Career advice usually treats "the ATS" as one machine. The vendors' own help pages describe at least three ways of matching, and the same employer can use more than one.

The exact-match rows are the ones that cost candidates interviews. On LinkedIn Recruiter, a recruiter who searches profiles for "customer service representative" gets that phrase and nothing else. A profile that says only "Customer Care Representative" does not contain it, and with no wildcard available, the recruiter would have to add the variant with OR. Greenhouse's Boolean help leaves word variants to the recruiter as well: the documented way to catch plurals and other forms is to add *. The article does not say plain terms are stemmed, so a search for "SOW" may or may not find "SOWs". Neither page mentions automatic synonyms.

The semantic rows are more forgiving. Greenhouse's Talent Matching extracts skills, job titles, years of experience and employment dates, then sorts applicants into Strong, Good, Partial or Limited Match. Greenhouse says it "does not automatically advance or reject candidates". How that ranking stage weighs your experience is covered in more depth in the AI resume screening guide.

You cannot tell from outside which method a given employer runs, or whether the recruiter types a strict Boolean string on top of an AI sort. So write for the strictest reader. If the ad says "statement of work", use those words; if your resume says SOW, write "statement of work (SOW)" once and the short form after that. A synonym-aware system loses nothing, and an exact-match search finds you.

Resume keywords by role, from employer job-posting data

Most "500 resume keywords" lists are written by hand. A better base for a list is the US Department of Labor's O*NET database, which tags software named in employer job postings. A "Hot Technology" is a software requirement among those most often found across all employer postings; an "In Demand" one turns up often in postings for that particular occupation.

The table below takes three things from each ONET occupation page, read on 13 October 2026: the job titles workers in that role report, the software flagged In Demand (where ONET marks any) and a selection of the Hot Technologies.

ONET flags Microsoft Excel for all six roles and Salesforce for five of them, so "Excel" in a skills list is expected rather than distinctive; it helps when the ad asks for it and you can show what you built with it. The title column matters more than most lists admit. ONET lists ten reported titles for customer service representatives alone, and an exact-phrase search for one of them misses the other nine.

Use the table as a starting set and let the job ad in front of you override it: where the ad says "Jira", write "Jira"; where it says "EHR documentation in Epic", name Epic. Certifications and licenses (PMP, your state RN license, a Salesforce administrator credential) belong in the same keyword set, because a recruiter can filter on them as easily as on a tool. How to lay these terms out in a dedicated section is covered in the resume skills section guide.

How to pull keywords from a job description: a worked example

Greenhouse's Project Manager ad from October 2026, for its Professional Services team, makes a good test case: a vendor that builds keyword search, hiring through its own product. Here is the process, step by step.

1. 1

   Copy the requirements and the duties into plain text

   Take the 'You should have' and 'What you'll do' lists. Ignore the mission statement and benefits; they carry no searchable requirements.

2. 2

   Mark each term required or preferred

   Everything under 'You should have' is required unless the ad says otherwise. 'PMP or another project management certification, a plus' is preferred.

3. 3

   Separate knockouts from keywords

   US work authorization and CST or EST business hours are eligibility facts that the application form asks about directly.

4. 4

   Write down the ad's exact wording

   Keep 'SOWs', 'RACI', 'NPS' and 'SaaS implementations' as written, and note which acronyms the ad never spells out.

5. 5

   Map each term to evidence

   For each required term, find the bullet where you did it. A term with no evidence stays off the resume.

Run on this ad, steps 2 and 4 give the following set.

The duties list adds data migration, solution design and change requests, which belong in bullets if you have done them. "3–5 years of customer-facing project management experience" is a threshold. Talent Matching extracts years of experience and employment dates from the resume, so make sure your dates parse rather than repeating "5 years" in your summary.

Placement follows the evidence. "Project Manager" goes in your headline if it is your title or an honest description of your role, because both exact-match search and Talent Matching read job titles. Acronyms get one spelled-out use: "statement of work (SOW)", "Project Management Professional (PMP)". The rest goes into bullets, where a recruiter reading the search snippet sees the term in context. An illustrative rewrite, using placeholder figures you would replace with your own:

• Before: "Managed customer projects and kept them on track."
• After: "Led [4] concurrent enterprise SaaS implementations with HRIS integrations and data migration; owned project plans, RACI charts and risk trackers, and managed change requests against each SOW to deliver on budget."

The second version covers seven of the ad's terms in one sentence a hiring manager would want to read anyway. To check your whole resume against an ad, the cvmakeover.ai ATS checker takes the job description with your upload, lists the ad's terms with the required ones flagged, and shows which ones your text contains. It counts a word and its plural as one match and accepts common abbreviations, which makes it more lenient than a strict Boolean search; its coverage figure is our own measure, not any employer's ranking.

Keyword stuffing, white text and other mistakes a recruiter sees

Stuffing fails on both kinds of system. Exact-match search shows the recruiter a snippet of the text around each hit, per Greenhouse's documentation, so a line of thirty comma-separated terms looks like what it is. Semantic matching gains nothing from repetition: Greenhouse's Talent Matching FAQ says several terms can map to one calibrated skill and "a longer list of matched terms doesn't always mean a higher match score".

Five other habits work against you.

• White or tiny text. Hidden keywords are still text. They come out in extraction, sit in the parsed profile and can appear in the search snippet beside your name, which is the worst place for a recruiter to find them.
• Pasting the whole ad. A skills list longer than the ad's own requirements reads as copying. Keep the terms you could defend in an interview.
• Skill bars and star ratings. A bar is a picture. The parser gets the word "Python" and nothing about the four filled dots next to it, so state a level in words ("advanced", "C1") if it matters.
• Only the acronym, or only the long form. "RN" alone misses a search for "registered nurse", and "Search Engine Optimization" alone misses "SEO". Use both once.
• Keywords with no evidence. A tool in the skills list that appears in no bullet gives the recruiter nothing to check. Talent Matching extracts skills from the resume as a whole, and the hiring manager reads the bullets.

Soft skills sit in a grey zone. Greenhouse's ad asks for "strong verbal and written communication skills", so the phrase is in the job's text, but nearly every ad has a version of it. Show it through a bullet ("led customer-facing steering meetings with VP-level stakeholders") rather than listing "communication" as a skill.

Resume keyword questions

What are ATS keywords?+

The terms a recruiter searches for, or an AI matching feature compares, in the text an applicant tracking system extracted from your resume. They are mostly job titles, tools, certifications, methods and degrees taken from the job ad.

How many keywords should a resume have?+

No vendor documents a target count, and Greenhouse says a longer list of matched terms doesn't always mean a higher match score. Cover each required term from the ad that you can back with evidence, plus the preferred ones you have.

Should I copy keywords exactly from the job description?+

Copy the wording, not the sentences. Exact-phrase searches on LinkedIn Recruiter and in Greenhouse's Boolean search match words as written, so 'customer care' does not find 'customer service'. Use the ad's term for skills you have, inside your own bullets.

Is 80% a good ATS score?+

Match percentages come from checker tools, and each computes them differently. None of the ATS documentation cited here describes showing candidates a percentage; Greenhouse's Talent Matching shows recruiters categories (Strong, Good, Partial, Limited Match). Treat a checker score as a list of gaps to fix.

Do keywords matter if a person reads my resume?+

Yes, for a different reason. A hiring manager scans for the same named tools and titles the ad asked for, and a bullet that names the tool and the result answers their question faster.

Key takeaways

• Keywords decide whether you appear in a recruiter's search results; knockout questions, not keywords, cause most automatic rejections.
• Greenhouse Boolean search and LinkedIn Recruiter match words as written, while Greenhouse Talent Matching, Workday Skills Cloud and Textkernel add synonyms. Writing for the strict case costs nothing on the lenient one.
• O*NET's job-posting data gives sourced starting lists by role, and its title variants show why the ad's exact job title belongs on your resume when it is true.
• Spell out each acronym once, keep the ad's wording, and put every claimed term in a bullet with evidence.
• Repetition and hidden text do not raise a semantic match score, and they show up in the search snippet a recruiter reads.