AI is now part of every recruitment conversation. But the conversation itself has moved on. Instead of asking if you use AI for hiring, we now talk about how you choose to use it. With more and more businesses choosing AI Applicant Tracking Software (AI ATS) for their recruitment, there are two clear models of usage emerging: automated AI ATS that make decisions about candidates or AI-assisted ATS that help teams better interpret data, while recruiters still make final decisions.
Some organisations are drawn to automated AI ATS because of budget, volume or time constraints. Others lean towards AI-assisted ATS because of their ethos, focus on candidate experience or desire to stay closer to each decision.
This article explores both approaches: what AI-assisted ATS platforms do today, how fully automated AI ATS differ, and how “human in the loop” will evolve – from checking every decision, to setting and reviewing the rules that guide our AI tools.
We’ll cover:
- What is an AI-assisted ATS and how does it differ from full automation?
- How do smart AI features enhance applicant tracking systems?
- Which tasks should AI support, and which should stay human?
- How will “human in the loop” change over time?
- Does using AI in recruitment affect candidate trust?
- Can AI applicant tracking systems support fairer hiring?
- What questions to ask before choosing an AI ATS
- The future recruiter: more human, better equipped
- What is Reach ATS’s philosophy on AI and Human in the Loop?
What is an AI-assisted ATS, and how does it differ from full automation?
As AI ATS become more prevalent, it’s important to understand both the capabilities and limitations of the technology. As previously mentioned, there are two clear models of AI ATS emerging:
AI-assisted ATS
- AI analyses information and presents insights
- Recruiters remain in charge of shortlisting, rejection and hiring
- Criteria, scoring and weightings are visible and can be altered
- AI provides data-backed evidence, humans make the decisions
This model is feature rich – it can summarise long CVs into short, readable candidate profiles, highlight key skills, qualifications or gaps against role requirements, and ranks candidates against agreed role criteria. It can also compare candidates within a cohort. Hiring teams see the summaries and suggestions, review the evidence provided and then decide how to proceed.
Fully automated or decision-making AI ATS
- The system itself acts on your behalf
- Prescribed rules or models reject or progress candidates automatically
- Recruiters decide which hiring steps to fully automate.
This model can save time for teams with ongoing high-volume hiring needs. But it does come with a level of risk attached. Criteria can be hard to view, change or explain. There is a chance that, left unchecked, bias can creep in. Decisions can also be harder to explain, potentially weakening candidate trust in the employer brand.
In simple terms, AI-assisted ATS informs the human; Automated AI ATS replaces human decision-making at prescribed stages.
Which tasks should AI support, and which should stay human?
To decide between AI-assisted ATS and fully automated ATS, consider the nature of the tasks:
Tasks AI can realistically automate today
AI is great at repetitive tasks that are based on patterns in text or data, and which are governed by clear criteria. Tasks it could handle might include:
Parsing and structuring information: turning CVs, cover letters and application forms into consistent fields, such as skills, years of experience, locations and sectors.
Highlighting key details: pulling out qualifications, certifications, languages and other facts that matter for a role.
Scoring against clear criteria: Applying weightings to skills, experience and answers that recruiters define, then suggesting a score or rank for each applicant.
Spotting matches and gaps: flagging candidates who are either a strong-fit, potential fit or clear mismatch based on must-have requirements.
Drafting content for review: Job descriptions, screening questions or candidate emails ready for human editing/approval.
All the above tasks use AI’s strength in reading and pattern-spotting but leave the recruiter in control. They support decision-making. In fully automated applicant tracking software some of these functions might also trigger instant actions, such as candidate rejection.
Tasks that need to remain under human control
Tasks that involve nuance, context or ethical choices, i.e. human judgement, should not be left to AI. For example:
Assessing culture and team fit: Fit is about behaviour, values and context, not just CVs. Humans are better able to weigh these factors as often they differ from team to team and role to role, requiring nuanced thought.
Handling exceptions: A candidate may lack a job requirement but bring rare experience that matters more. HR teams are much better able to balance the job requirements with these exceptions. There are no criteria you can set for these exceptions!
Balancing fairness and risk: Employment gaps and non-linear career paths can easily lead to automated AI ATS rejection. Humans look for the story behind the data.
Rejecting or progressing candidates: Of course, for organisations hiring at volume, automated AI can help speed up the shortlisting process by rejecting candidates who don’t meet essential criteria. But that also risks great candidates slipping through the net. AI should flag low-scoring candidates, but the final hiring decision should always sit with a human.
How will “human in the loop” change over time
Right now, human in the loop usually means that a person reviews AI insight for each role and candidate – the system suggests and the recruiter decides.
But as AI continues to improve, our role in the loop might shift. Rather than approving every decision, teams might decide to focus more on setting and reviewing the criteria the AI uses.
In practice that might mean:
- Agreeing the rules and weightings for scoring candidates
- Choosing which parts of the process can/should be automated
- Checking sample cases to see how the system behaves
- Changing criteria when strategy, law or workforce needs change.
It’s worth remembering that “human in the loop” doesn’t necessarily mean a person touches every single record. It means that people remain in control of what the AI is allowed to do, and how it judges/analyses candidates. Automated AI ATS lean more toward this future shape (with risk attached as detailed previously). AI-assisted ATS are a way of building towards a more “hands off” approach, while keeping case-by-case control in the here and now.
Does using AI in recruitment affect candidate trust?
Candidates are not blind to AI. Many are already using it themselves to write their CVs and prep for interviews. It would be fair to say that many assume some type of automated screening is happening behind the scenes. Some applicants are comfortable with this. Others are worried about being rejected by a system they cannot see or challenge.
The type of AI you use has a direct effect on trust.
AI-assisted ATS give teams a strong, evidence-based story to tell. It’s easy to explain how AI is used to organise information and aid parity and consistency, while a human reviews applications, and makes final decisions.
A fully automated system can feed a sense that hiring is ambiguous and unfair. If the system has rejected them, teams may not be able to explain why in clear terms, making it harder to provide reassurance.
There are of course ways to be more open and transparent about your use of AI. For example, you can:
- Update your privacy notice to explain where and how AI is used in your recruitment process
- Add a line in application acknowledgement emails outlining how AI tools are used to review applications (and if using AI-assisted ATS, that people will always make the decisions)
- Train hiring managers to answer questions about AI with simple, honest explanations.
Transparency won’t fix every concern, but it shows respect to candidates. It signals openness and honesty.
Can AI applicant tracking systems support fairer hiring?
Fairness is one of the main reasons people are interested in using AI for recruitment. And it’s also one of their greatest AI concerns. Used well, AI can help reduce certain types of bias. For example:
Consistent Criteria: Using AI to score candidates removes personal preference from early screening.
Focus on Skills: AI can highlight skills and experience rather than details (names, postcodes etc.) that might lead to bias. Supporting more equitable treatment.
Reporting: AI applicant tracking highlights drop off patterns for different groups, giving you evidence to review your stages and content for fairness.
It’s important to remember that AI is not neutral by default. Models learn from data; if past hiring decisions were biased, AI tools will not only repeat the bias but potentially amplify it. AI can support fairer hiring. But it does not remove recruiter’s responsibility to design fair processes and act on evidence.
What questions to ask before choosing an AI ATS
If you’re ready to pick an AI applicant tracking system, then look beyond the features list to really get to grips with its core functionality and how it will impact your team’s hiring. Try asking:
Who makes the final decision on progression and rejection?
Can my team turn off automated rejection rules or override AI suggestions?
Can we see and change the criteria used for scoring?
How easy is it to adjust weightings for skills, qualifications or experience? Can you tailor them per role?
How does the system explain its recommendations?
If a candidate is ranked first, can the recruiter see why? Is there a clear link between criteria and score?
What data does the AI use and how is it governed?
Are models trained on your data, general data, or both? How are privacy, retention and audits handled?
How does the platform support fairness and inclusion?
Does it offer anonymised screening, structured scoring and outcome reports?
What training and support is provided?
AI is only useful if people know how, when and why to use it. Does the vendor offer training and ongoing support?
Above all, be on the lookout for impossible claims. AI cannot “replace recruiters” or “remove all bias”. Look for clear explanations on decision-making, data, governance and candidate rights. Choosing any ATS, AI-assisted or fully automated, is as much about values and trust as it is about features.
The Future Recruiter: more human, better equipped
There is concern within HR that AI will make recruiters less important. The reality is likely the opposite. As AI ATS take on more of the repetitive hiring work, the value of human skills will increase. Recruiters oversight will be crucial to ensure processes stay fair, clear and human. They’ll use the improved data insights to guide hiring strategy and of course, their unique human capabilities to spot the nuance and stories behind CVs and candidate scores.
What is Reach ATS’s philosophy on AI and human in the loop?
Reach ATS was designed to automate repetitive manual hiring admin; freeing teams to focus on finding the right talent for their business. And that hasn’t changed.
Our smart AI ATS offers structured data and simple summaries that provide the information you need at a glance. It supports scoring and ranking based on criteria you control, and that you can explain.
Right now, our AI tools do not automatically reject candidates on your behalf. We firmly believe that an AI-assisted ATS approach is best. You decide who progresses or leaves the process.
But we also recognise that this relatively new technology is moving quickly, and that different organisations will make different choices based on their ethos, risk appetite and budget. We will continue to explore fully automated AI in low-risk hiring areas, but will remain true to our guiding AI principles:
- Humans will always set and review the criteria AI uses.
- There will always be clear controls and explanations.
- Key decisions that affect people’s lives will remain accountable to people.
Our aim is simple: give hiring teams intelligent assistance inside a modern, human-centric ATS today, while leaving space for even smarter automation tomorrow.
If you’d like to see our AI-assisted, intuitive applicant tracking system in action, then Reach out and book a demo today.