Control · Human in the loop
AI sources, the recruiter decides
Human-in-the-loop recruiting means an AI agent does the repetitive work while people approve the steps that matter and make every decision about candidates. In AIcruiters your team sets approval rules for search strategy, candidate selection, outreach and handover. Low-confidence cases always go to a recruiter, and warm candidates are handed over with full context.
What does human in the loop mean in recruitment?
The term describes where a person sits in an automated process. In the loop: the system prepares an action and waits for a person to approve it. On the loop: the system acts on its own, and a person monitors and can step in. Out of the loop: nobody checks. For recruitment, the third option is not acceptable. Whether a step should be in or on the loop depends on the risk of the step and how much the team trusts the output.
So AIcruiters has no single on/off switch for automation. Control is set per action. A search strategy for a new client and a follow-up message to a candidate who already showed interest carry very different risks, so they deserve different rules. The platform overview shows where each step sits in the workflow.
Which actions can you put behind an approval?
Approval rules cover the four points where an agent's action has a real effect on candidates or clients. For each, your team decides whether it needs approval or may run automatically.
- Search strategy: check the requirements and titles from vacancy intelligence before searching
- Candidate selection: decide who moves from the ranking to outreach
- Outreach: approve or edit messages before they are sent
- Handover: decide how and to whom warm candidates are passed
- Search strategyNeeds approval
- Candidate selectionRecruiter reviews shortlist
- OutreachNeeds approval
- Follow-upAutomatic
- Low confidenceAlways to a recruiter
How strict should the rules be?
An illustration of how a team can move from strict to lighter control. The right setting depends on your roles, clients and experience with the agent.
What happens when the agent is uncertain?
Uncertainty is routed to a person, not resolved by a guess. Typical cases are a profile with conflicting work history, a candidate who meets the hard requirements on paper but whose experience is hard to interpret, or a reply that is neither a clear yes nor a clear no. These cases go to a recruiter together with the reason for the review.
This rule does not loosen as trust grows. Even when a team lets most steps run automatically, low-confidence cases and unexpected outcomes still go to a recruiter. How rankings show their evidence, and why missing information is flagged instead of scored as a failure, is explained on the candidate ranking page.
What does the recruiter receive at handover?
When a candidate responds positively, the agent stops and a recruiter takes over. The handover contains what is needed to call the candidate straight away, without reconstructing the history.
- The candidate profile
- The match rationale: why this person was selected for this vacancy
- The full conversation so far
Why does this matter for the EU AI Act and the GDPR?
The EU AI Act classifies AI systems used to evaluate candidates as high-risk (Annex III, point 4) and requires such systems to be designed so that people can effectively oversee them (Article 14). Under the GDPR (Article 22), people have the right not to be subject to a decision based solely on automated processing that significantly affects them, with limited exceptions.
Human-in-the-loop design is how you make that oversight real rather than a checkbox: the recruiter sees the evidence, can override the outcome and decides who is invited, assessed and hired. This is a general explanation, not legal advice. Read our guide to the EU AI Act in recruitment and the security and compliance page for more.
Questions about human-in-the-loop control
How much control do recruiters keep?
You decide which actions the agent may carry out independently and where approval is required, for searches, candidate selection, outreach and handover.
Does AIcruiters make hiring decisions?
No. It identifies and prioritises candidates and starts conversations. Interviews, assessment and the hiring decision stay with your recruiters and hiring managers.
Can we change the rules later?
Yes. Rules are configuration, not a one-time setup. A sensible approach is strict approval at the start, adjusted as you review the output.
Does approval slow everything down?
Reviewing a prepared search strategy or message is quicker than building or writing it yourself. And the approval points are exactly where recruiter judgement adds the most.
What if a recruiter disagrees with the ranking?
They override it. The evidence behind each score is visible, so the recruiter can see where the agent's reading differs and adjust the criteria.
Set the rules for your own team
In the demo we walk through the approval points on a real vacancy, and discuss which steps your team wants to keep in its own hands.