Quick answer
The four major ATS platforms handle your resume differently. Greenhouse never auto-rejects or scores — every application reaches a human who prioritizes by keyword search and scorecards. Lever matches word stems but handles acronyms poorly, so spell terms out. iCIMS builds its own skills profile from your full resume text. Taleo can auto-reject via knockout questions and requisition-level scoring before a human looks. What works across all four: clean single-column structure and exact role-specific keywords.
Same pipeline, four different rulebooks
Every ATS runs the same basic pipeline: parse your resume into structured data, add it to a searchable index, let a recruiter search and filter. Jobscan's analysis of job descriptions from over 12,000 companies found Greenhouse used by 19.3% of them, Lever by 16.6%, and iCIMS by 15.3% (Workday leads even this broader dataset at 15.9%, on top of the Fortune 500 dominance covered in our Workday guide). All four of the platforms below are genuinely common, not niche. If you're weighing which ATS resume optimization service can actually help with parsing quirks like these, we've compared the best-known ones honestly.
What differs, and differs a lot, is what happens after parsing: who scores you, how forgiving the keyword search is, and whether a bad match can auto-reject you before a human ever opens your file. Those differences change what's actually worth optimizing for on each one.
Greenhouse: no robot rejects you, but a human scorecard does
Greenhouse, used by roughly 7,500 companies including Airbnb, DoorDash, Stripe, and Dropbox, is explicit that it does not auto-reject or algorithmically score resumes. Greenhouse CEO Daniel Chait has said outright that any automated scoring system "is subject to the biases of the people who are building the algorithm," and that scoring is better left to hiring managers. Every rejection in Greenhouse is a deliberate human decision.
What Greenhouse does instead is feed your parsed data into a scorecard built around Focus Attributes: specific skills and traits a hiring manager defines when the job is posted, then assigns to individual interview stages so each interviewer grades only their piece (a phone screen might focus on culture fit, a technical round on a named skill). The practical move is to read the job description for its exact nouns and verbs, since those are very likely the literal labels on that internal scorecard, and to write resume bullets as mini evidence statements (situation, action, result) a recruiter can lift straight into their notes.
Lever: word-stemmed search that still can't read acronyms
Lever (LeverTRM), used by more than 7,000 companies including Netflix, Spotify, and Shopify, doesn't score or rank your resume either. That's left entirely to the recruiter reading it. But Lever's own search is more linguistically flexible than most: it supports word stemming, so a recruiter searching "collaborate" will also surface resumes containing "collaborating," "collaborated," or "collaborative."
That flexibility has a hard limit. Lever does not expand abbreviations or acronyms. A resume that only says "SEO" won't surface for a recruiter searching "search engine optimization," and a degree listed as "B.S." won't surface for a search on "Bachelor of Science," or vice versa. Since Lever doesn't rank you either, the entire game is simply appearing in the search results at all, so spell out both the acronym and the full term for every credential and skill where you have the space.
iCIMS: an AI Role Fit score built from your whole resume, not your Skills section
iCIMS runs recruiting for more than 6,000 companies, including 40% of the Fortune 100 (Microsoft, IBM, Target, UPS, and Uber among them). Two things make it distinct. First, iCIMS auto-generates its own list of a candidate's skills from the full text of the resume, not from whatever you typed into a dedicated Skills section, which means burying a skill only in your bullet points still gets it captured, but so does the reverse: a Skills section stuffed with terms that never appear in your actual experience can read as thin.
Second, iCIMS applies its own AI scoring, called Role Fit, that groups candidates into tiers a recruiter can scan at a glance. By iCIMS's own documentation, a candidate's ranking is job-specific and based on "experiences and skills that match the requirements of the job," so the same resume can land in a different tier for two different postings at the same company, which is exactly why tailoring per job still matters even with an AI layer doing some of the sorting.
Taleo (Oracle Cloud): the one that can reject you without a human involved
Oracle Taleo and its successor Oracle Cloud HCM run recruiting for 57 Fortune 500 companies (UnitedHealth Group, JPMorgan Chase, Ford, Kroger, and Goldman Sachs among them) and just over 5% of all companies in Jobscan's broader database. It's the one platform on this list with a confirmed, documented auto-reject capability. Oracle's own workflow tools let recruiters configure automatic disqualification for candidates who don't complete a required test, lack a stated certification or degree, or fall below a threshold on Taleo's AI-driven Suggested Candidates score (which rates Profile, Education, Experience, and Skills on a 0-3 star scale each). Because every employer configures these workflows differently, there's no way to know from outside whether a given Taleo instance has auto-reject switched on. That makes completing every application field, not just the resume upload, a real requirement rather than a nice-to-have.
Taleo's own keyword search is also the least forgiving of the four. In Jobscan's testing, it doesn't recognize tense, plural, or abbreviation variants at all. A search for "project manager" won't return "project management," and a search for "Certified Public Accountant" won't return "CPA." The fix is blunt but effective. Write out both forms of every meaningful credential and title, every time you have room.
The pattern underneath all four
Every one of these platforms runs the same three-step pipeline we cover in how ATS screening actually works: parse, index, search. What changes, sharply, is who does the scoring (a human scorecard, an AI tier, or a literal keyword match with no scoring at all) and whether a bad match can quietly disqualify you before anyone reads your file. Generic "make it ATS-friendly" advice misses all four of those differences, which is exactly why we research and write for each platform, and each industry's specific vocabulary, rather than shipping one template and calling it universal.
Where we come in
We rewrite your resume and LinkedIn profile so both parse cleanly and surface for the exact keywords recruiters in your field search — verified by an expert ATS reviewer and delivered in 72 hours.