How applicant tracking systems work

Short answer

An ATS is the workflow. Parsing is one service inside it.

A submitted file may be stored, converted into text, parsed into fields, and attached to a candidate profile. A recruiter can then review the original document, search fields or resume text, apply filters, use a matching feature, or move the applicant through configured stages.

Pipeline

From file to recruiting workflow

This diagram is a general model. Products can combine stages, skip them, or use a third-party parser.

01 / INPUTResume filePDF, DOCX, RTF, TXT, or another accepted type

02 / GATEFile ingestionUpload checks, storage, malware scan, size and type rules

03 / EXTRACTDocument extractionText, content order, metadata, OCR when available

04 / PARSESection + entity parsingJobs, employers, dates, education, contact data, skills

05 / MAPCandidate profileNormalized fields plus the original attachment

06 / RETRIEVESearch + matchingKeyword search, filters, optional similarity features

07 / PROCESSRecruiting workflowReview, stages, questionnaires, rules, interviews, disposition

Terms

Six concepts people collapse into “the ATS”

Related systems and decisions
Concept	What it does	What it does not prove
Applicant tracking system	Stores applications and supports recruiting stages, records, communication, and reporting.	That every customer enables the same modules.
Resume parser	Turns document content into fields such as employer, title, dates, and skills.	That the candidate is qualified.
Recruiter search	Finds candidates using text, structured fields, saved criteria, or combinations.	That missing one word always rejects an application.
Knockout question	Captures a required answer and can trigger a configured rule.	That the resume itself received a low score.
Matching or ranking	Compares available candidate data with job criteria using product-specific logic.	That another employer or checker uses the same model.
Recruiter decision	Moves, rejects, contacts, or reviews a person inside the workflow.	That the action was automatic.

Document layer

Extraction comes before understanding

A PDF can store text in a content order that differs from its visual order. A DOCX stores paragraphs, runs, and tables in XML. An image-only PDF may need OCR before software can search its words.

When extraction is wrong, later stages inherit the error. A job date can drift away from its employer. A contact icon can disappear. Two columns can interleave.

Open the parser notes →

candidate_profile.jsonparsed

{
  "name": "Jordan Lee",
  "currentTitle": "Data Engineer",
  "employers": ["Northstar Labs"],
  "skills": ["Python", "SQL"],
  "sourceFile": "resume.pdf"
}

Product variation

Real systems expose different controls

The same employer can also connect outside assessment, matching, and parsing products.

Greenhouse

Documents resume-to-profile parsing, recruiter keyword search, and auto-reject rules tied to configured application answers. Those are distinct features.

Greenhouse parsing docs

Workday

Documents resume-derived candidate fields and an optional Candidate Skills Match feature with Strong through Low labels.

Workday skills match docs

Lever

Documents a parser that adds fields such as name, organization, and contact information to the candidate summary.

Lever parsing docs

Oracle Taleo

Oracle documentation states that Taleo Enterprise resume parsing is delivered through a third-party partner service.

Taleo documentation

Decision logic

How can an application be rejected?

CONFIGURED

Application rules

A disqualifying answer can trigger a defined rule. Greenhouse documents this for custom application questions.

HUMAN

Recruiter review

A recruiter can decline an application during review and record a reason in the ATS.

CONNECTED

Assessment or integration

An outside screening or matching product can return information used in the workflow.

VARIABLE

Match feature

A configured product may calculate fit labels or grades. The model and customer setup matter.

Practical controls

What you can actually test

01
File acceptance
Use the requested extension and stay within the upload limit.
02
Text extraction
Search, select, and copy the final document. Read the pasted order.
03
Field clarity
Keep each title, employer, date, and set of bullets together.
04
Truthful alignment
Use the posting’s terms where they accurately describe your work.

Run the local checks Review ATS myths

References

Sources used on this page

Unsuccessful resume parseGreenhouse Support
Documents parse failures, a 2.5 MB parsing limit, and layout-related errors.
Auto-rejectGreenhouse Support
Shows that configured application-question responses can trigger automatic rejection.
Talent Matching data processing FAQGreenhouse Support
Lists resume fields extracted for an optional matching feature.
Concept: Candidate Skills MatchWorkday Documentation
Documents an optional skills-match feature, its supported files, and its Strong to Low labels.
Prospects and CandidatesWorkday Documentation
Explains resume-derived fields and the configurable Candidate Skills Cloud feature.
Understanding resume parsingLever Help Center
Describes the fields Lever adds to a candidate profile after parsing.
Getting Started: Career Section overviewOracle Help Center
States that Taleo Enterprise resume parsing is delivered through a third-party partner service.
Resume Parser FieldsRChilli Documentation
Shows the structured fields a commercial parser can return, including experience, education, contact details, and skills.
Structure of a WordprocessingML documentMicrosoft Learn
Documents the paragraph, run, and text structure inside DOCX files.
What you may be missing when you search PDF documentsPDF Association
Demonstrates how PDF content order can differ from the order visible on the page.