What is an applicant tracking system (ATS) and how does it actually work?

An applicant tracking system, almost always shortened to ATS, is software that manages hiring from the moment a job opens to the moment someone is hired [2][4]. It stores every application in one database, turns each resume into a structured profile, and tracks candidates through the stages of a workflow, so a recruiter, a hiring manager, and an interviewer all work from the same record instead of a scatter of email threads and spreadsheets [2][4].

The category started in the 1990s as digital filing for resumes and now covers most of what a recruiting team touches: posting jobs, collecting and screening applications, scheduling interviews, storing feedback, and reporting on the funnel [2][4]. Adoption is close to universal at large employers, where Jobscan detected an ATS at 97.8% of Fortune 500 companies in 2025, or 489 of the 500 [1]. For a job seeker, that means an application to a big company usually lands in an ATS first, where it is parsed and sorted before a human opens it [1][4].

What sits inside the system

Most systems share the same parts under the marketing: a place to create and approve a job, a way to push it to boards and a careers page, an intake that captures applications and parses resumes, a screening layer that filters or ranks candidates, a pipeline that shows where everyone stands, and a reporting layer for metrics and compliance data [3][4]. Vendors differ mainly in how much automation and AI they wrap around those parts, but the skeleton is consistent across Greenhouse, Workday, Lever, and Workable [3][6][7].

Why teams adopt one

Coordination is the real driver, more than any single feature. Hiring puts several people on the same candidates over weeks, and a shared record keeps the application, notes, feedback, and current stage in one place that updates as people act, which is what stops candidates from stalling unseen [2][5].

A job enters an ATS as a requisition, the formal request to fill a role that routes for approval before anything goes live [6]. Once approved, the system publishes the opening, collects applications into one database, and parses each resume into a structured profile so the pool is searchable from the first submission [3][4].

Requisition and approval

The requisition is the record every later step attaches to, and it usually carries the job description, the hiring team, the budget or headcount sign-off, and the pipeline candidates will follow [6]. Enterprise systems put real workflow here; Workday supports creating requisitions on mobile and evergreen requisitions that stay open for continuous hiring [6]. This step decides which openings are genuine and funded before any applicant sees them [6].

Posting and distribution

Posting pushes the approved job to a company careers page and out to external job boards, often to many at once from a single action. Workable publishes a job to more than 200 boards simultaneously, including Indeed and LinkedIn, and BambooHR connects postings to boards like Indeed, ZipRecruiter, and LinkedIn [3][8]. Distribution sets the size of the funnel, since a role on one board draws a narrower pool than one syndicated across many [3].

Application capture

Capture is where applications from every channel collect in one place instead of separate inboxes. Candidates apply through forms on the careers page or a job board, and the ATS records each one against the requisition, often through a branded, mobile-friendly form with auto-fill and one-click options that reduce drop-off [3][4]. Centralizing capture replaces forwarded resumes with a list the whole hiring team can see and sort [4].

Resume parsing

Parsing turns an unstructured resume into data. The system reads the file and extracts contact details, work history, skills, and qualifications into a standardized profile, so no one retypes anything and applications compare side by side [2][3]. Workable reports parsing a file in roughly three to five seconds and pre-filling fields like name, email, phone, and experience [3]. Parsing quality is uneven across formats, which is why heavily designed resumes with tables or columns sometimes arrive with garbled fields, a limitation covered later [3].

Screening narrows a large pool to a shortlist using application questions, keyword and criteria matching, and increasingly AI scoring [2][4]. It is also the stage with the most legal exposure, because a filter that disproportionately excludes a protected group can create liability under Title VII even when no one intended it [10].

Knockout and screening questions

Knockout questions automatically filter out candidates who miss a hard requirement such as work authorization, a required license, or minimum experience [4]. They come as yes/no, multiple choice, or short text, and a disqualifying answer can move an applicant out of contention without human review [3][4]. Used narrowly for genuine minimum qualifications they save reviewer time; set too aggressively they discard people who could do the job, which is why the questions themselves deserve scrutiny [4].

Keyword and criteria matching

Beyond hard filters, an ATS ranks or flags candidates by matching resume content against the job's requirements. Older systems compare keywords and qualifications, while newer ones use semantic matching that reads for related skills rather than exact words and shows a match score; Lever offers AI resume review that flags best-fit applicants, and Workable generates candidate summaries with match scores [3][7]. A score is a suggestion built from patterns in the data, not a verdict on whether someone can do the work [3].

Automated scoring and its limits

AI scoring is now common and carries specific compliance risk. The U.S. Equal Employment Opportunity Commission issued technical assistance in May 2023, titled Assessing Adverse Impact in Software, Algorithms, and Artificial Intelligence Used in Employment Selection Procedures Under Title VII, warning that an automated selection tool can produce disparate impact regardless of intent [10]. The guidance treats the four-fifths rule as a rough screen for adverse impact rather than a definitive test, and it states that an employer can be liable even when an outside vendor built or administers the tool [10]. That leaves the testing and recordkeeping duty with the employer, and none of this is legal advice [10].

Once candidates are in the system, an ATS manages them as a pipeline: a set of stages each person moves through, with a shared database, scheduling, feedback, and messaging attached to the record [4][7]. This is the part recruiters work in day to day, since it shows every open role and where each candidate stands [7].

The candidate database and pipeline stages

The pipeline is a visual board of hiring stages, and candidates advance or get archived as decisions are made. Stages typically run from applied through phone screen, interview, and offer; Lever displays a Pipeline by Stage across sourced, applied, screen, interview, offer, and hired, and Greenhouse builds fully configurable stages such as application review, recruiter screen, and hiring manager review with a visual pipeline showing how many sit in each [4][5][7]. That shared view keeps candidates across several roles from stalling unseen [5][7].

Scorecards and structured feedback

Structured feedback has interviewers rate candidates against defined criteria instead of leaving freeform impressions. Greenhouse is built around structured hiring: teams define competencies and interview kits for each job, then complete standardized scorecards so decisions rest on consistent evidence, and Lever auto-generates scorecards and summaries from AI interview transcripts [5][7]. Standardizing feedback makes candidates comparable and lessens the pull of a single strong first impression [5].

Interview scheduling

Scheduling coordinates candidate availability, interviewer calendars, and rooms or video links, one of the most time-consuming parts of hiring to do by hand. Workable offers candidate self-scheduling with two-way calendar sync to Google and Outlook and video links for Zoom, Meet, and Teams, and Ashby can automatically generate recommended schedules for high-volume funnels or complex panels [3][9]. Automating this removes rounds of back-and-forth email that otherwise slow every hire [3][9].

Candidate communication

Communication keeps the process moving and stays logged against each record. An ATS centralizes emails, notes, and documents and connects to email and calendars so messages and interview invites live with the candidate rather than in a personal inbox, and Workday additionally routes feedback through Microsoft Teams and Slack while giving candidates status updates [4][6].

Reporting turns pipeline activity into metrics, and compliance features collect the demographic and recordkeeping data hiring laws require [4][11]. Both run off the same central record, so every application, stage change, and hire is captured as data the moment it happens [4].

Recruiting metrics and source tracking

An ATS reports on the hiring funnel: how many candidates entered, how they moved between stages, and how long each step took, through measures like time to fill, time to hire, and qualified candidates per hire [4]. It also tracks source of hire, showing which boards or channels produce candidates who actually get hired so budget can shift toward what works [3][8]. Greenhouse, Lever, and Ashby all treat pipeline and source analytics as core, with dashboards for pipeline health and stage-to-stage conversion [5][7][9].

EEO and OFCCP compliance data

Compliance features exist because U.S. hiring is regulated, and the ATS is where the required data gets collected and stored. Many systems include EEO reporting that helps employers meet equal employment opportunity obligations [4]. Federal contractors carry an added rule: the OFCCP Internet Applicant Rule defines who counts as an applicant, generally someone who applied electronically, was considered for a specific role, met the basic qualifications, and did not withdraw before an offer, and it requires those contractors to invite applicants to self-identify race, ethnicity, and gender before an offer on a voluntary basis and to keep the records [11]. That definition also scopes adverse-impact analysis, since the reported pool is built from these records and the EEOC's guidance on automated tools turns on whether selection rates differ across protected groups within it [10][11]. An ATS handles the solicitation and retention that would be hard to do by hand across thousands of applications, and none of this is legal advice, so contractors confirm current obligations with counsel [11].

An ATS manages candidates through a single hiring process, a recruiting CRM nurtures relationships with people who have not applied yet, and an HRIS manages employees after they are hired [2][7][8]. The three overlap at the edges and increasingly ship together, which is why the labels blur in practice [7][8].

ATS versus recruiting CRM

A recruiting CRM stores and warms up potential candidates before there is an application, while an ATS manages active applicants against specific open roles. The difference is timing: the CRM works the top of the funnel with sourcing and nurture campaigns, and the ATS takes over once someone applies [7]. Many platforms bundle both, so the line is often internal; Lever markets LeverTRM as ATS plus CRM, and Ashby lists CRM and sourcing alongside its ATS, letting a candidate be nurtured and then tracked in one tool [7][9]. Comparing an ATS to CRM software, a common shorthand, reflects that shared database lineage [2].

ATS versus HRIS

An HRIS, or human resources information system, is the system of record for employees: payroll, benefits, time off, and performance, everything after a hire, while an ATS handles everything before it. The handoff is the tell. In BambooHR the ATS is one module inside a broader HR platform, and a hired candidate's data transfers automatically into Employee Records to start onboarding [8]. Workday works the same way at enterprise scale, with recruiting inside the wider Workday HCM suite so a candidate becomes an employee record without re-entry [6].

Where the lines blur

Large suites now cover all three functions, so a buyer cannot read a product's category from its name. Workday and SuccessFactors are HCM suites that include recruiting and lead the Fortune 500 partly because hiring, HR, and payroll sit together, while standalone tools like Greenhouse and Ashby stay focused on hiring and connect to a separate HRIS through integrations [1][5][6][9]. The practical question is whether recruiting, sourcing, and employee data need to live in one system or connect across several [1][9].

An ATS makes hiring coordinated and measurable, but it introduces failure modes worth watching: imperfect parsing, over-tuned filters, biased automation, and the overhead of running it well [3][4][10]. The tool organizes a process; it does not by itself guarantee good hiring decisions [5].

Parsing and format sensitivity

Resume parsing is reliable for clean, standard layouts and less reliable as formatting gets complex. Multi-column designs, tables, and graphics can produce garbled or missing fields, so a strong candidate can look weak in the parsed profile through no fault of their qualifications [3]. Some systems read .docx more accurately than PDF and quality varies by vendor, so teams that rely on parsed data should spot-check it rather than trust it blindly [3].

Filters that cut too deep

Knockout questions and keyword screens save time but can quietly remove qualified people. A filter that demands an exact phrase or a specific credential drops candidates who describe the same skill differently or hold an equivalent qualification [4]. The loss is invisible, because the recruiter never sees who was screened out, which is why screening criteria deserve periodic review against who is actually reaching interviews [4].

Automated screening and legal exposure

AI ranking can introduce or amplify bias, and the liability sits with the employer. The EEOC has warned that an automated tool producing adverse impact can violate Title VII even when a third party built it, and that a vendor's assurance does not remove the employer's duty to check for disparate impact [10]. Teams using AI screening are expected to test for adverse impact and keep records, making this a compliance question as much as a technical one [10][11].

Cost, setup, and data hygiene

An ATS only reports well if the team uses it consistently, and a system full of half-updated records produces unreliable metrics. Setup takes real work, since stages, scorecards, permissions, and job board connections all need configuration, and enterprise suites like Workday carry heavier implementation than lighter tools [6][9]. The value comes from disciplined use across the team rather than the software alone [5].