TL;DR
- An applicant tracking system does three jobs: collects applications in one place, moves candidates through named stages, and records why each decision was made
- Four federal Acts govern discrimination in Australian employment, covering age, disability, race and sex, and state or territory law applies on top
- Packaged Australian systems combining tracking with onboarding start at $75 per user per month excluding GST, plus $350 up front to set the account up
- The third job, the audit trail, is the one employers undervalue and the one that matters if a decision is ever questioned
Short version: the filtering is what everyone argues about. The written record of why you filtered is what you will need.
Job seekers widely believe an applicant tracking system is a robot binning CVs that lack the right keywords. Small employers widely believe it is enterprise software they do not need.
Both are roughly half right, and both get the same half wrong. What the system is really doing is creating a record.
The three jobs an ATS does
Strip out the marketing and that is it. Applications from every channel land in one queue. Candidates sit at a named stage, so anyone can see where a person is without asking. Every movement between stages is logged with a reason attached.
Australian buyers looking at Talent Onboard HR CRM systems and similar packaged builds will find candidate pipeline management, interview scheduling and assessments, offer letter and contract handling, and onboarding in one product, at $75 per user per month on Professional and $125 on Enterprise, with $350 up front per account. The Enterprise tier adds quoting and invoicing, which agencies billing clients per placement need and internal HR teams generally do not.
Built on Zoho CRM by an Australian Zoho partner, that one aims at both recruitment agencies and HR managers running internal hiring. The more useful configuration, if you would rather not run two systems that disagree with each other.
The screening question, honestly
Keyword filtering exists. It is cruder than vendors imply and less absolute than candidates fear. Most systems rank rather than reject, and a human decides. The problem is not that the ranking gets used. It is that the ranking is often the only thing recorded.
| What the system can capture | Whether it should drive a rejection |
|---|---|
| Required licence, ticket or registration | Yes, if the role genuinely requires it |
| Right to work in Australia | Yes |
| Years of experience stated | Cautiously, as a signal not a gate |
| Keyword density against the ad | No, as a decision on its own |
| Employment gaps | No |
| Postcode, name, age, photo | No, and preferably do not collect it |
The bottom row is where automated screening turns into a problem. Not because a system chose to discriminate, but because a proxy did it invisibly and nobody wrote down why the candidate was moved to rejected.
The legal ground underneath it
Australian employers operate under four federal discrimination Acts covering employment: the Age Discrimination Act 2004, the Disability Discrimination Act 1992, the Racial Discrimination Act 1975 and the Sex Discrimination Act 1984. The Australian Human Rights Commission’s quick guide to Australian discrimination laws sets these out alongside the state and territory legislation applying in every jurisdiction, and makes the point that an exemption under one Act does not carry across to another.
This is not legal advice and a lawyer should be advising on any specific decision. The software consequence is simple enough to state, though. A hiring system that records a stage change and nothing else leaves you reconstructing your reasoning from memory, months later, under pressure.
Testing works the same way. Assessments are useful and widely used, and the risks are well covered in this guide to using pre-employment assessments without creating legal risk. The recurring theme in both areas is that defensibility comes from consistency, applied to every candidate the same way, and evidenced.
The overview below shows tracking and onboarding sitting in one pipeline, which is the arrangement that keeps the record intact after the offer is accepted.
Video: Talent Onboard Streamline Recruitment & HR in One Smart System (watch on YouTube).
Configure these three things on day one
Give every stage a required reason. If moving a candidate to rejected does not force a selection from a fixed list, people leave it blank and the record is worthless.
Restrict the fields. Do not collect what you should not consider, because the easiest way to stop a photograph influencing a decision is not to have one in the file.
Set a retention rule. Unsuccessful applicants are people whose personal information you now hold, so decide how long you keep it, set it in the system, and do not leave it to whoever tidies up.
Automation, and the limit of it
Voice agents are arriving in high-volume recruitment. Retell AI style agents handling first-pass screening calls and writing structured notes into the record do relieve pressure on teams drowning in applicants, and the structured notes are the useful part more than the call handling.
Where it stops being appropriate is any point where a judgement about a person gets made rather than information gathered. A machine can record that a candidate holds the required licence. Deciding someone is not right for a team is a human act, and the file should show a human doing it.
What good looks like six months later
Pull a candidate at random from last quarter’s rejections. See whether you can explain, from the system alone and without ringing anyone, exactly why they did not progress.
If you can, the system is doing its job. If you cannot, you bought a queue.
