When AI takes over part of recruitment, what Is left of the recruiter’s role?

AI is now automating many of the tasks that once accounted for a significant part of recruitment. Sourcing, matching, writing, pre-screening and administrative tasks can already be largely supported by AI. The real question, therefore, is no longer simply, “What can AI do?” but rather, “Where does the recruiter still add value when technology performs part of the work?”

Screening Is not the same as evaluating

A matching tool can compare a CV with a job description, identify relevant skills and rank candidates. When dealing with large volumes, the time savings are significant. But is a score enough to assess a candidate?

A system works on the basis of the objectives, criteria and data it is given. If these are incomplete, biased or poorly defined, it can reproduce the same mistake with remarkable efficiency. The International Labour Organization highlights precisely this risk: some HR systems rely on overly narrow objectives, imperfect data or opaque processes.

It can, of course, be argued that human assessment is also subject to bias, and that is true. One of the benefits of automation is precisely its ability to apply criteria more consistently and, when a system is properly designed, to make those criteria testable and documentable. But this strength can also become a weakness: while a recruiter may make an error when assessing one candidate, a poorly designed automated criterion can reject hundreds of profiles in exactly the same way.

Take a simple example. If five years of experience are defined as a mandatory threshold, even though four years spent working on a highly comparable project would actually be more relevant, the system may apply the wrong rule perfectly.

The recruiter’s role therefore shifts. It becomes less about personally carrying out every stage of the screening process and more about understanding what the tool is measuring, why it is measuring it and when its output needs to be challenged.

Understanding what the client really needs

This ability to assess and interpret begins even before candidates are reviewed. A client asking for “a senior profile with five years of experience” is almost never expressing the entirety of the underlying need. Behind the job description there may be a team that needs reassurance, a project that is still poorly defined, a low tolerance for risk or a previous experience that has shaped the manager’s expectations.

AI can analyse a brief, suggest questions and identify certain inconsistencies. But it does not have the history of the relationship or the contextual knowledge that enables a recruiter to determine which questions really matter, and sometimes to challenge the stated requirement when it does not correspond to the actual problem.

The distinction is therefore no longer simply between what “humans can do” and what “machines cannot do”. It increasingly lies in the professional’s ability to interpret context, exercise judgement and take responsibility for that judgement.

When candidates are using AI too

This evolution does not affect recruiters alone. Candidates are also using AI to improve their CVs, prepare for interviews, draft answers or complete certain assessments. According to Indeed’s Global AI Survey, 70% of the jobseekers surveyed already use generative AI tools, including to research companies, write cover letters and prepare for interviews. As a result, the apparent quality of an application is gradually becoming a less reliable indicator of actual competence.

The more AI improves the signals visible to recruiters, the more recruiters need to learn to look beyond them: exploring an experience in greater depth, asking why a particular decision was made, presenting the candidate with a real-life situation or checking whether they can clearly explain what they claim to know.

At the same time, automation should not make the candidate invisible within the process. It can also lead to more decisions that are difficult to understand, including rejections made with little or no genuine human interaction.

The GDPR already addresses this risk. Article 22 provides, under certain conditions, the right not to be subject to a decision based solely on automated processing where that decision produces legal effects or similarly significantly affects the individual. It also provides safeguards, including human intervention and the possibility of challenging certain decisions.

In other words, better recruitment automation also means preserving transparency and ensuring that candidates are able to understand the process to which they are being subjected.

The human relationship remains difficult to automate

Assessment is, moreover, only one part of recruitment. Recruitment also depends on something far more difficult to reduce to data or a score: the relationship itself.

A chatbot can schedule an appointment, answer certain questions and even formulate persuasive arguments. But persuading a professional to leave a stable position, understanding an unspoken hesitation or supporting someone when the process becomes difficult requires an ongoing relationship built on trust. Recruiters also put their own credibility on the line when recommending a company to a candidate or a candidate to a client.

Automation can therefore produce two opposite effects: it can replace human contact and make the process more impersonal, or it can free recruiters from administrative work and allow them to spend more time on interactions that genuinely add value.

The challenge, therefore, is to define clearly what AI can take over and what should remain the recruiter’s responsibility.

Towards the augmented recruiter

This shift in where value is created has direct consequences for the skills recruiters will need. Recruiters will, of course, still need to understand the roles they are hiring for and to assess candidates effectively. But they will also need to know how to interpret a matching score critically, identify the criteria that produced it, question the limitations of a tool or provider, verify its output and document important decisions.

As discussed in our previous article on the evolution of jobs and skills, AI proficiency is therefore becoming a professional skill in its own right.

Article 4 of the AI Act, which has applied since February 2025, already requires providers and organisations deploying AI systems to take measures to ensure an appropriate level of AI literacy among the people using those systems on their behalf.

For AI systems used in certain high-risk areas, including recruitment and employment, the timetable is different. Following the EU Digital Omnibus regulation, the specific obligations applying to systems covered by Annex III will apply from 2 December 2027.

The two should therefore not be confused: AI literacy is already a current obligation and business priority, whereas part of the specific regulatory framework governing high-risk recruitment systems will apply at a later stage.

And what about recruiters’ jobs?

There remains one final, and perhaps more uncomfortable, question. If a recruiter equipped with AI can process more applications, automate part of the sourcing process and spend less time on administrative work, a company may indeed need fewer people to handle the same volume of activity.

That does not necessarily mean that the profession will disappear. It is more likely to mean that recruitment teams will evolve: less value will be placed on manually processing large volumes, and more on advisory work, assessment, relationship-building and effective use of technology.

Early use cases already provide an indication of this shift. LinkedIn reports, for example, that users of its Hiring Assistant save more than four hours per role and review 62% fewer profiles. The question, therefore, is not whether productivity gains exist, but what organisations choose to do with them.

Reducing headcount is one possibility. Using the time saved to improve assessment, provide clients with better advice and offer candidates a better experience is another. That is probably where the real transformation of the recruitment profession will take place.

At Triglav Digital, this is how we see the role of AI in recruitment: not as a substitute for the recruiter’s judgement, but as a tool that should enable recruiters to spend more time on what genuinely creates value for both candidates and clients.

AI can produce an answer.

The professional remains responsible for knowing when to accept it, and when to challenge it.

To continue exploring this topic, you can also read our previous article on how AI is transforming jobs and skills. And if these developments are raising questions within your organisation, our teams would be pleased to discuss them with you.

Rym, from Triglav Digital.

References

1. BERG, Janine & JOHNSTON, Hannah. AI in Human Resource Management: The Limits of Empiricism. ILO Working Paper 154, International Labour Organization, 11 November 2025.
2. INDEED. The Indeed Global AI Survey: Your Guide to the Future of Hiring. Survey conducted among more than 7,000 jobseekers and HR professionals.
3. EUROPEAN UNION. Regulation (EU) 2016/679 — General Data Protection Regulation (GDPR). Article 22 on automated individual decision-making, including profiling.
4. EUROPEAN COMMISSION. AI Literacy — Article 4 of the AI Act. Applicable since 2 February 2025.
5. EUROPEAN UNION / EUROPEAN COMMISSION. AI Act — Application Timeline Following the Digital Omnibus. Rules concerning high-risk AI systems covered by Annex III apply from 2 December 2027.
6. LINKEDIN. How Can Businesses Close the AI Adoption Gap? Start with Your Recruitment Teams. LinkedIn Research, October 2025.