skip to content

AI-led interviews: the underestimated risk in executive hiring

AI-led interviews: the underestimated risk in executive hiring
contents

An interview conducted by artificial intelligence is no longer an experiment. According to Greenhouse's 2026 report on AI interviews (opens a new window), conducted among 2,950 candidates in the United States, the United Kingdom, Ireland, Germany and Australia, 63% of US candidates have already been interviewed by an AI, 13 points more than six months ago. The same report shows the other side of this rapid adoption: 38% of US candidates have already dropped out of a hiring process because it included an AI interview.

For a company hiring in volume, that figure looks like a loss of efficiency. For a portfolio company looking for a CFO, a COO or a business unit head in the months after a deal, it becomes a value creation risk. The profiles that withdraw first are often those with the most options.

What the figures actually say

Before drawing conclusions, the data needs to be read carefully.

  • A phenomenon measured first in the United States. The figures cited here relate to the US sample (1,200 respondents), according to Greenhouse's press release (opens a new window). The French market is not covered, but practices usually reach it with a lag of a few quarters.
  • Dropout that has already happened, not just an intention. The 38% are candidates who actually withdrew. A further 12% say they would.
  • A massive transparency deficit. For 70% of candidates, the use of AI had not been clearly announced before their last interview of this kind. And 21% only found out once the interview had started.
  • Silence after the effort. Of the candidates who went through an AI interview, 51% never received a reply.

In other words, the problem does not come from the technology alone. It comes from the way it is deployed.

The four reasons candidates walk away

The Greenhouse report identifies the main reasons for withdrawal:

  • No human involved: 33% cite pre-recorded video interviews, scored with no person present.
  • No transparency: 27% mention that the use of AI had not been announced.
  • Surveillance: 26% mention the feeling of being watched and analysed by a tool.
  • Obligation: 26% refuse an AI interview imposed on them with no alternative.

These four reasons have one thing in common: they concern the relationship, not what the tool does. A candidate accepts being assessed. They are less willing to be assessed without knowing by whom, against what criteria and with no possibility of dialogue.

Why the risk is higher for executive roles

For executive roles, the hiring dynamic is reversed: it is often the company that has to do the convincing.

  • The best profiles are not actively looking. An executive approached for a CFO or COO role is weighing up several approaches. An impersonal first contact is enough to make them disengage.
  • The process is read as a governance signal. A senior candidate infers from it how the company treats its teams, its clients and its decisions. A mechanised process projects the image of an organisation that delegates what should be a matter of judgement.
  • The cost of a missed hire is high. In a value creation plan, an unfilled key role delays the roadmap for the first 100 days, the financial structuring or the rollout of a new offering.
  • Reputation travels in a small circle. Experienced executives talk to each other, to their advisers and to the funds. A bad experience as a candidate quickly spreads beyond the company concerned.

The question is therefore not whether portfolio companies should use AI in hiring, but where it creates value and where it destroys it.

The regulatory framework is tightening

The legal dimension raises the stakes for European companies.

  • The European AI Act classifies AI systems used for hiring and candidate selection as high-risk systems (Annex III of Regulation (EU) 2024/1689 (opens a new window)). That classification brings obligations on transparency, human oversight and risk management.
  • The GDPR already regulates decisions based solely on automated processing (Article 22 (opens a new window)), with a right to obtain human intervention.
  • Candidates' expectations point the same way: 57% of US candidates believe companies should be legally required to disclose their use of AI.

Transparency is therefore not just good employer-brand practice. It is becoming a compliance requirement.

How to avoid dropout: five practical principles

The Greenhouse data also shows that AI is not doomed: when the interview goes well, 38% of US candidates come away with a better image of the employer. It all comes down to execution.

  • Announce the AI before the interview, not during it. State in the invitation which tool is used, what it assesses and who makes the final decision. This is information that 44% of candidates explicitly ask for in advance.
  • Offer a human alternative: 46% of candidates want to be able to ask for an interview with a person. For executive roles, this option should be the rule.
  • Keep AI for the stages where it genuinely adds something. Initial qualification at high volumes, logistics, summarising files: yes. First contact with a targeted executive, assessment of leadership, judging suitability for a fund and a team: no.
  • Guarantee feedback every time. No candidate who has given time to an interview should be left without an answer. It is the point where the gap between current practice and expectation is easiest to close.
  • Formalise a policy on the use of AI in hiring. Only 18% of US candidates believe employers have a clear policy.

Conclusion

The AI-led interview meets a real need for speed and volume. But the figures show it drives away a significant share of candidates when it is imposed, opaque or left without follow-up. For executive roles, where every profile counts and the hiring process puts the company's credibility on the line, the issue goes beyond the tool.

The right approach is to distinguish what can be automated from what calls for judgement. AI can speed up the process. It does not replace the conversation that persuades an executive to join a project.

FAQ

Should you give up on AI in hiring?

No. AI is useful for handling volume, organising stages and structuring information. The risk appears when it replaces human contact at the decisive stages, particularly for executive roles.

What does European regulation say about AI in hiring?

The AI Act classifies hiring and selection AI systems as high-risk, which implies, among other things, transparency and human oversight. The GDPR also guarantees the right not to be subject to a decision based solely on automated processing.

take action

turn reading into hiring

describe your role, our specialised headhunters take it on. 26 days on average.