What is an applicant tracking system (ATS)?

What an ATS is, in one paragraph

An applicant tracking system, usually shortened to ATS, is the software a company uses to keep every open job and every candidate for it in one place. It holds the job, the applications, each candidate's CV and contact details, the stage they have reached, the notes your team wrote about them, and the interviews booked with them. That is the whole idea. When a colleague asks who applied for the office manager role last Tuesday, or which of the three developers you met is still waiting on an answer, the system answers without anyone opening a spreadsheet or scrolling back through an inbox.

Everything else grows out of that core. Some systems read the CV and fill in the profile for you. Some publish a careers page so applications arrive already inside the system instead of as attachments. Some book the interviews. The common part is always the same: one record per candidate, one pipeline per job, and one place the whole team looks.

What it replaces

Most small teams do not start with an ATS. They start with three things that each work well on their own and badly together.

The first is the spreadsheet. It has a tab per role and a row per candidate, and it is accurate on the day it is written. A week later two people have edited it in different windows, someone has sorted a column without selecting the others, and nobody trusts the "status" field any more.

The second is the shared inbox. Applications arrive as attachments, which means the CV lives in an email thread and the notes about the candidate live in a reply to that thread, unless they live in a direct message to a colleague. Finding out what happened to a candidate means reconstructing a conversation.

The third is the calendar back-and-forth. You email three possible times, the candidate replies two days later that none of them work, you send three more, and by the time a slot is agreed the candidate has finished a process somewhere else. This is the part that costs the most time and the part most teams accept as unavoidable.

An ATS replaces all three with one record. The CV, the stage, the notes and the interview sit against the candidate's name, and anyone on the team can see the same thing.

What happens to a CV inside an ATS

A CV enters the system either because a candidate applied through your careers page or because a recruiter uploaded the PDF. What happens next is called parsing: the system reads the document and pulls out the structured facts. Contact details. Links to a LinkedIn or GitHub profile. The current job title and company. The skills. Good parsing also produces a short summary a recruiter can read in ten seconds instead of opening the PDF.

Those facts become a profile, and the profile enters a pipeline. A pipeline is the ordered list of stages a candidate moves through for one job: applied, screened, interviewing, offer, hired, and whatever else your process needs. Moving a candidate from one stage to the next is the single most common action in any ATS, and the pipeline view is the screen most recruiters live in.

Candidates who are good but not right for this job do not disappear. They stay in the talent pool, which is simply every profile you have collected, searchable. Six months later, when a similar role opens, the pool is the first place to look, and a role you have filled once is cheaper to fill the second time.

Do you need one at five hires a year?

Honestly, it depends on what those five hires cost you in hours rather than on the number itself. A team hiring five people a year with one recruiter, no other stakeholders, and candidates who answer emails within the hour can run on a spreadsheet for a long time.

The point where that stops working is usually not volume. It is people. As soon as two or three colleagues need to see the same candidate, or a hiring manager wants to know where things stand without asking you, the spreadsheet becomes a thing you maintain rather than a thing you use. The other trigger is repeat hiring. If you fill the same kind of role twice a year, the value of a searchable talent pool is real, because the second search starts from the leftovers of the first.

Do the arithmetic once. Count the hours your last hire lost to scheduling and re-typing CVs, multiply by what an hour of your time is worth, and set that against a month's subscription. For most small teams the scheduling alone clears the fee by the second hire.

What "AI" means in an ATS today

The word is on every vendor's homepage and it usually covers three concrete things.

The first is enrichment. The system reads an uploaded CV and fills the profile: contact details, links, current title and company, a skills list. This is the most useful kind, because it removes typing rather than judgment.

The second is summarizing. A recruiter-ready paragraph that says who this person is and what they have done, so a hiring manager gets the gist without a PDF.

The third is generating text, such as a careers page written from your company website.

What AI should never do is overwrite a recruiter. If you typed a candidate's title yourself because you learned it in a call, no model should replace it because the CV says something else. The right behavior is narrow and boring: fill the blanks, leave everything a person entered alone. Any tool that quietly rewrites your work is costing you more attention than it saves. It is worth asking a vendor this directly, because the answer tells you how much they trust their own model.

Interview scheduling is where most ATSs stop

Here is the gap that surprises people. A system will happily store the candidate, parse the CV, move them through five stages and record the outcome, and then, at the moment you need to actually meet them, hand you back to email. You leave the tool, write "what does Thursday look like for you", and wait.

That is the expensive part of hiring, and it is the part a system can genuinely remove. The mechanics are not complicated. A recruiter sets their weekly interview hours once. Each candidate gets a personal link showing only the slots that are genuinely free, because the events already in the recruiter's Google Calendar block the rest. The candidate picks one, with no account and nothing to install, and the interview lands in the recruiter's calendar with a Google Meet link attached. The same link is what they use if they need to move it.

If you want the detail on how to set that up, the guide on scheduling interviews faster walks through it step by step, and the interview invitation email template covers the message that carries the link.

A short glossary

These are the words you will meet in every demo, in the meaning the industry actually uses.

Term	What it means
Pipeline	The ordered set of stages a candidate moves through for one job.
Stage	One step in that pipeline, such as screened, interviewing or offer.
Talent pool	Every candidate profile you have collected, searchable across jobs.
Parsing	Reading a CV file and extracting structured fields from it.
Enrichment	Filling a profile from parsed data and public links.
Careers page	The public page listing your open jobs, where applications arrive.
Time to hire	Days from a candidate entering the pipeline to accepting an offer.
Time to interview	Days from deciding to meet someone to actually meeting them.
No-show rate	The share of booked interviews a candidate does not attend.

The last two are worth watching even if you track nothing else. They are the two numbers a small team can move in a week, and they are the ones candidates feel.