Applicant Tracking System: What an ATS Does and How to Choose

An applicant tracking system (ATS) is software that stores job applications and helps a team track candidates through its hiring process. You use it to see who applied, what stage each application has reached, what evidence reviewers collected, and who owns the next action. Features such as resume parsing, scheduling, automated screening, and AI assistance depend on the product and its configuration.

If you are choosing your first ATS, begin with a real role and a real hiring process. A polished dashboard cannot show you whether a candidate can submit an application, an interviewer can find their assignment, or you can recover your records when you leave.

This guide explains the category, separates commonly confused features, and gives you a practical evaluation and rollout plan.

What does ATS mean in recruitment?

ATS stands for applicant tracking system. Its central job is to organize recruitment records and the work associated with them. Oracle’s Talent Acquisition Cloud glossary describes the category in terms of handling recruitment electronically; the term itself does not promise any particular AI capability.

Think of the ATS as the shared record for a hiring decision. A candidate may apply to several roles, so the person and each application need to remain distinguishable. The same person can be unsuitable for one opening and a strong fit for another.

For your team, the important questions are concrete:

• Which applications need a first review?
• Which candidates are waiting for an interview or an answer?
• What did each reviewer observe?
• Who can make the next decision?
• What should the candidate hear next?

A spreadsheet can answer some of these questions. The case for an ATS grows when several people work with the same records and your current process loses ownership, context, or follow-through.

How does an applicant tracking system work?

An ATS connects a role, its applications, and the steps your team uses to evaluate people. The exact workflow varies, but a useful process makes the current state and next action visible.

Here is an illustrative hiring flow, not a required sequence:

SAP’s official recruiting overview provides a concrete product example covering applications, candidate statuses, interviews, and feedback. Treat it as an example of possible capabilities. Check what the specific product and plan you are evaluating actually support.

The software does not decide whether your process makes sense. If every stage asks the same questions, an ATS can faithfully organize repeated work. Define what each stage needs to establish before configuring it.

How is an ATS different from a CRM or HRIS?

The useful distinction is the main record each system organizes. An ATS centers on applications to roles, a recruitment CRM centers on relationships with potential candidates, and an HRIS centers on employee administration. Products can combine these functions.

Do not buy based only on category labels. An ATS might include sourcing tools, or require an integration. An HR suite might include recruiting, but its application flow still needs testing. A job-board connection might mean publishing, importing, or merely linking to a posting.

Write down the task you need completed, then ask the vendor to demonstrate that exact task.

Which ATS features should a small team evaluate first?

Start with the features needed to complete your hiring process reliably. Advanced automation matters less if the team cannot find applications, submit feedback, or answer candidates.

A usable candidate record

Check whether you can find the original application, attachments, correspondence, stage history, and relevant feedback together. Verify how the system handles one candidate applying to different roles. Ask what happens when a candidate corrects their email address or submits missing information.

Clear ownership and structured feedback

Assign an owner to each step and define what they need to record. Your ATS should help reviewers produce useful evidence rather than a collection of unstructured opinions. Our guide to structured interview scorecards explains how to make that feedback comparable.

A workable candidate journey

Apply from a phone, upload a real document, and follow every confirmation and scheduling link. Test errors as well as successful submissions. Check whether a candidate can contact a person when the process does not work for them.

Appropriate access and an exit path

Test the actual permissions of a recruiter, interviewer, and administrator. Then request an export and inspect its contents. A downloadable list of names is different from an export that preserves applications, notes, attachments, and relationships.

Remove a test interviewer from the role and try previously shared links again. Ask how account access, attachments, integrations, and shared links are revoked. A hidden navigation item is not enough evidence that access has ended.

Understand where candidate data goes

Ask where the records are hosted, which providers receive them, and whether any AI feature uses applicant information for model training. Review the relevant terms instead of assuming every feature uses the same data path. Agree on retention and deletion handling for your process, then test what happens to a sample record, its attachments, and any linked integration. These are evaluation questions; the applicable obligations depend on your arrangement and jurisdiction.

For U.S. employers, the Department of Justice’s guidance on AI and disability discrimination in hiring explains why using hiring technology does not remove the need to evaluate disability-related barriers. Accessibility and accommodation handling belong in the evaluation, with legal requirements assessed for your jurisdiction.

Does an ATS automatically reject resumes?

An ATS can support automatic rejection, but resume parsing, searching, screening rules, and AI assessment are different operations. You need to know which mechanism is enabled before drawing conclusions about an application outcome.

Greenhouse’s parse-failure documentation explains that a resume can remain attached when its information cannot be filled into profile fields automatically. Its separate auto-reject documentation describes rules triggered by answers to application questions. These are documented behaviors in that product, not universal rules for every ATS.

During a demo, submit an application that triggers each configured rule. Check the resulting status, history, notification, and recovery path. If a tool makes recommendations, ask what data it uses and what a reviewer can inspect before acting.

For the candidate-side implications, see our deeper explanation of ATS auto-rejection myths. If AI is central to your buying decision, use the separate AI-native ATS evaluation guide.

How should you test an ATS before buying?

Run a short working session using your own hiring scenario. Ask the vendor to complete the whole process, including an exception and an export, rather than showing isolated features.

Use fictional candidate data or volunteers who understand the test. Avoid uploading real applicants’ information to several trial systems just to compare interfaces.

For each result, record whether it was demonstrated, requires an integration, requires manual work, or is unavailable. Record the plan and configuration used. A capability described on a roadmap should not count as available for your current process.

Ask about costs after identifying the required configuration. Include seats, add-ons, integrations, implementation, support, and data migration where relevant. The separate guide to ATS renewal and switching costs can help you prepare that conversation.

How do you implement an ATS without copying a broken process?

Define your decision criteria first, configure the smallest useful workflow, and pilot it on a role your team understands. Avoid importing years of inconsistent records before you know how the new process should work.

The U.S. Office of Personnel Management’s job-analysis guidance connects job tasks, competencies, and assessment. Its federal context differs from startup recruiting, but the methodological question is useful: what work must this person do, and what evidence would show they can do it?

Use that question to drive the rollout:

1. Agree on the role. Record the work, requirements, and decision owner.
2. Map the stages. Give each stage a purpose, reviewer, and expected output.
3. Prepare communications. Explain the process and provide a route for questions or accommodation requests.
4. Set permissions. Give people access needed for their responsibilities and test it.
5. Migrate deliberately. Decide which records are necessary, check duplicates, and verify attachments and application relationships.
6. Run a pilot. Follow applications through the process and correct confusing steps before expanding.

Write the process down so it does not depend on one person’s memory. A new recruiter should be able to identify why a stage exists, who owns it, and what happens when the usual path fails.

How do you know whether your ATS is helping?

Compare the process before and after adoption using consistent definitions. Measure work and waiting time in your own team rather than assuming a vendor’s headline improvement will apply to you.

Useful local measures include:

• Time from application submission to first meaningful review.
• Applications waiting beyond your team’s agreed response window.
• Feedback assignments completed before the decision meeting.
• Candidate messages that needed correction or manual chasing.
• Missing or incorrectly linked records found during migration checks.

These are suggested measures, not industry benchmarks. Set your own review window and record why it fits the role and team. Keep candidate experience alongside operational speed: moving someone quickly to an unexplained rejection is not the same as running a useful process.

When a metric changes, inspect the cause. A smaller applicant pool, a different role, or a paused budget can change hiring time independently of the software.

Common questions about applicant tracking systems

Do you need an ATS for your first hire?

There is no universal hiring-volume threshold. A simple shared process may work while one person owns the search. Consider an ATS when handoffs, permissions, feedback, and candidate follow-up become hard to manage consistently.

Is an ATS the same as AI recruiting software?

No. An ATS can organize hiring without model-based assessment. AI features may assist with particular tasks, but you still need to inspect their inputs, permissions, outputs, and failure behavior.

Can an ATS replace a recruiter?

The system can organize records and automate supported tasks. Your team still needs to define the role, evaluate evidence, make decisions, and communicate responsibly. Ask which work becomes easier and which responsibilities remain with people.

Will switching systems preserve all your data?

Only an inspected export and migration test can answer that for your setup. Check attachments, stage history, feedback, custom fields, and links between candidates and applications. Confirm what can be imported into the destination product before closing the old account.

How can you try this hiring workflow in Kit?

Start with one approved role and configure the application and assessment stages your team needs. Kit organizes job postings, candidate applications, interviews, work samples, code assignments, and team review; select the stages that fit the work rather than adding every available option.

Keep budget authorization and other systems’ responsibilities clear. A hiring pipeline records how you evaluate candidates; it does not by itself approve spending or replace payroll and employee administration.

Test whether your team can follow an application, collect relevant evidence, make a decision, and tell the candidate what happens next. You can start a Kit trial and run that test with a sample role before committing your live process.