Coming from Workday Recruiting? See the 12-step escape guide → Read it
Stage 3 · 8 min read read · Last reviewed 2026-05-23

What Is an ATS and How Does It Work? Resume Parsing, Pipeline Stages, and Screening Explained

An ATS (Applicant Tracking System) is software that companies use to receive, organise, and manage job applications. It’s the system on the other side of the “apply now” button.

This guide explains how ATS tools work from the inside — specifically how resume parsing, pipeline stages, and keyword screening operate in practice across the major tools.

What happens when you click “Apply Now”

When a candidate submits an application through an ATS:

  1. The application lands in the ATS database — not in an email inbox. Every application is stored as a structured record associated with a specific job requisition.

  2. The resume is parsed — the ATS attempts to extract structured data from the resume file (PDF, Word, text): name, email, current title, employment history, education, skills. This extraction is automated and imperfect.

  3. The candidate record is created — all parsed data populates a candidate profile within the ATS. This profile follows the candidate through the entire hiring process.

  4. The candidate appears in the recruiter’s queue — typically in a “New” or “Applied” stage, sorted by application date or by a match score (depending on how the recruiter has configured screening).

  5. The recruiter reviews and takes action — advance to phone screen, reject with or without a message, or tag for future consideration.

How resume parsing works

Resume parsing is the automated conversion of your unstructured resume into structured ATS fields. Every major ATS does this:

What the parser extracts:

  • Contact information (name, email, phone)
  • Current job title and employer
  • Employment history (company names, titles, dates)
  • Education (institutions, degrees, graduation years)
  • Skills (extracted from keywords found anywhere in the document)

Where parsing fails:

  • Two-column or table-heavy resume formats (the parser reads columns left-to-right, merging unrelated content)
  • Text in image files (can’t be read at all)
  • Skills or contact info in sidebar text boxes (often skipped)
  • Non-standard date formats (parsers expect “January 2022 – March 2024” or similar)
  • PDF files with unusual encoding or heavy graphics

The parsed output is what recruiters see as your “candidate profile.” If your resume parses badly, your candidate profile will have gaps, errors, or missing sections — and the recruiter may not look beyond it.

How pipeline stages work

ATS tools organise candidates through a defined sequence of stages. A typical pipeline in Greenhouse or Workable looks like:

  1. Applied — application received, not yet reviewed
  2. Phone Screen — recruiter call scheduled or completed
  3. First Interview — hiring manager or team interview
  4. Technical / Skills Assessment — role-specific evaluation
  5. Final Interview / Onsite — late-stage interview
  6. Offer — offer extended
  7. Hired / Rejected — terminal stage

Companies configure these stages differently. A startup might have 3 stages; an enterprise might have 8. Every stage change is logged with a timestamp, which creates the hiring audit trail.

Candidate view of stages: Most ATS tools send automated email notifications when a candidate is moved between stages. “We’ve reviewed your application and would like to invite you to a phone screen” is typically an automated trigger from a stage change, not a personally written email.

How keyword screening works (and what it doesn’t do)

There is significant misunderstanding about how ATS keyword screening works. The reality varies by tool:

Greenhouse: Does not automatically reject candidates based on keywords. Keyword filtering is manual — recruiters can search their pipeline by keyword, but applications are not auto-rejected based on missing keywords. Greenhouse presents all applications; the recruiter decides.

Workable: Has an AI match score that ranks candidates against job spec keywords. A low match score doesn’t auto-reject, but it deprioritises a candidate in the recruiter’s view — they see high-match candidates first.

SmartRecruiters: Has configurable screening questions and knockout criteria. Companies using SmartRecruiters can set automatic rejection rules (e.g., “if the candidate answers No to ‘do you have the right to work in the UK?’ auto-reject”). This is opt-in configuration, not default behaviour.

Lever and Ashby: Primarily manual review — recruiters see all applications and filter by keyword search or tags rather than automated scoring.

The honest answer: Auto-rejection happens most at enterprise ATS tools where volume makes it necessary (SmartRecruiters, iCIMS), and primarily via screening questions (work authorisation, minimum qualifications) rather than keyword absence. Most SMB ATS deployments rely on human review for all but the clearest screening criteria.

The recruiter’s actual workflow

Understanding what a recruiter’s ATS screen looks like helps candidates:

  1. Application list view — a table of candidates with columns: name, application date, current title, match score (if enabled). Recruiters scan this quickly, looking for titles and companies that signal fit.

  2. Candidate profile view — the parsed resume plus any custom screening answers, notes from previous conversations, and stage history. Recruiters typically spend 20–40 seconds on this view before making an initial advance/reject decision.

  3. Bulk actions — recruiters can select 50 applications and reject them in one action. This is how high-volume roles get processed when 400 applications arrive for 1 open position.

  4. Keyword search — recruiters search for specific skills or tools (“Python”, “HMRC”, “Figma”) across all candidates in a pipeline. Your skills section determines whether you appear in these searches.

What this means for candidates

The first screen is fast. In high-volume roles, the first human review of your application is 20–40 seconds. The parsed title and most recent employer are the primary signals. Skills extracted from your text are secondary.

Parsing quality is your responsibility. A clean, single-column, plaintext resume minimises parsing errors. The ATS doesn’t know your resume had an elegant two-column design; it just sees merged, out-of-order text.

Screening questions matter more than keywords in most tools. If a role has a “do you have the right to work in the UK?” screening question and you answer No, you will be automatically rejected before a human sees your CV. Answer all screening questions carefully and honestly.

Your stage history follows you. If you apply to the same company multiple times (different roles), the ATS links your records. Recruiters can see that you applied before and how far you progressed. A strong previous application is a positive signal; a previous rejection for the same role with the same profile is a negative one.


Further reading

Go Deeper