More employers are now using an AI agent, rather than a person, for first-round interview calls. The AI agent asks structured questions on a voice or video call, follows up on answers, and hands the hiring team a scored transcript.

The appeal is clear for a company swamped with applications and wanting to automate using AI. However, for Senior and Leadership candidates with in-demand skills, interest from other companies and possibly a counteroffer already on the table, the positioning of an AI-led interview looks very different.  

So how do top-flight engineers feel about being interviewed by an AI Assistant? The headline is that candidates need an interview to be a two-way conversation. They have questions that require thoughtful, detailed answers - on the culture, values, leadership style, organisational structure etc - to be able to assess their fit with a role and company.

For these reasons, highly qualified candidates may also be those most put-off by an AI-led interview. Indeed, many will take this as a red flag and evidence that a company is not the right one for them.

Here’s what the evidence on AI interviews says, sentiments from engineers we’ve spoken to, and where we think it is heading.

Startups and Scaleups

Startups have broadly adopted AI in recruiting, but less common among startups are fully AI-conducted first-round interviews, one-way recorded video interviews scored by AI or AI-only candidate evaluation without human interaction.

The reason is straightforward: startups care as much about motivation, adaptability and cultural fit as technical competence. Founders need one or more conversations with candidates who will be crucial to product development and the growth of the company, rather than outsourcing that judgement to an AI agent.

Interviewed By AI

Some candidates report feeling more relaxed on an AI-led interview, without a human interviewer and often also without a digital avatar present. There is no judgement of appearance, facial expressions or assessment of body language (these features were dropped due to legal scrutiny and bias concerns) and there is more flexibility to schedule the interview 24x7 which is useful for candidates in different time zones.

There are different types of fully autonomous AI-driven interviews - where an AI system generates questions, listens or reads responses, dynamically generates follow-ups, and produces an automated evaluation or score against specific criteria. These include: chatbot screening, asynchronous pre-recorded video, live two-way voice interviews, live technical and coding interviews, and simulations and role play assessments.

These differ from AI-assisted interviews, where a human conducts the interview and the AI transcribes the dialogue, tracks keywords, monitors interview compliance, or provides structured candidate summaries.

How the systems are set up by employers is crucial. A digital rubric is used to establish what a ‘high’, ‘medium’ or ‘low’ quality answer will sound like for each question. NLP is then used to transcribe the audio or video responses into text, analyse and score them to give a final rating. A cut-off score is used to automatically filter candidates.

The AI systems are very structured, every candidate is asked the same questions and scored by the same rubric. This should reduce bias but it doesn’t allow the candidate to ask questions or for a meaningful conversation. With over 35 years’ experience of interviewing, we know that every candidate is different and conversations flow in different directions.

The Candidate Backlash

Overall candidates object less to AI itself than to being assessed by it without warning, explanation of what is being scored, a human in the loop or any feedback after the interview.  

How employers are preparing and using AI to represent them in interviews is where the real problems lie.

• Trust is low. The backlash is mostly about trust and a lack of transparency. 82% of UK jobseekers report they were not told beforehand, and 24% only realized AI was involved after the interview had started. (Greenhouse)
• Seniority is hard to score. The co-founder of one AI voice-interview start-up told IEEE Spectrum that its tool has mostly been tested on entry-level roles. Senior hires are judged on decision making, cultural fit, leadership and influence, which in our view are hard to capture in a scripted conversation.
• It feels one-sided. Greenhouse's chief people officer said that candidates feel "processed rather than considered". It is a filtering exercise rather than a conversation, a test of whether a candidate can include the keywords and concepts that the AI is scoring them on.
• Candidates expect more. They want guardrails: disclosure, an explanation of what is being measured, feedback, and the option of a human interviewer.

Although there are no published figures for Senior candidates alone, all this matters most for senior and lead engineers because they are the candidates with choices; a process that feels disrespectful is an easy way to lose the people an employer needs most.

The Drop-Out Problem

Candidates who drop out are likely to come away with a negative view of the employer, damaging the company’s brand. A senior engineer may be put off immediately by an AI interview, the absence of a human moderator and the opportunity to get answers to their questions.

The cost of getting it wrong is high. An employer rarely sees the candidate who left and for scarce candidates, one drop-out costs far more than for a high-volume role.

Overall, in the UK, 30% of candidates have already quit a hiring process because it involved an AI interview. This figure is likely to be much higher for more senior candidates.

Will AI-Led Interviews Improve?

AI-led interviews are here to stay but we expect the way they are used to improve. Candidates are walking away from badly run AI - crucially those parts of hiring it really should have no place - rather than from AI as such. Disclosure, adjusting the style of communication and prompt feedback are all fixable.

Indeed, companies could now be fined by the Information Commissioners Office (ICO) for using AI interviewing tools without transparency or safeguards. In March 2026 the ICO found that many UK employers describe their tools as decision support when, in practice, no one meaningfully reviews the outcome. Employers relying on solely automated decisions must tell candidates, offer human review and allow them to challenge the result.

Even with these improvements it is unlikely that AI-driven interviews will be suitable for senior hires or startup and scaleup companies.

Some tech companies have now banned the use of AI during interviews or are moving back to human-to-human video interviews and in-person technical tests to combat cheating by candidates who are also using AI during interviews to generate responses in real time.

Meta and Google are two tech giants who do not use AI agents to lead interviews but are allowing candidates to use an AI coding assistant integrated into the interview CoderPad environment to test AI fluency.

The Human Layer

The point of rigorous screening and interviewing by a highly experienced human is to select a handful of right-fit, qualified candidates. When RWA interviews and assesses candidates and provides a shortlist, an employer can shorten its own recruitment process.  

Fewer stages, a human in the loop from the first point of contact with a candidate, no costly drop-outs due to unanswered questions. Counter-offers, notice periods, salary band queries and second thoughts are handled by people.

None of this means avoiding AI. We use it ourselves for search, and every hour it saves on admin is an hour for the conversations that decide whether a placement works.

If you are hiring Senior Software Engineers and want to talk through the balance, book an exploratory, no-obligation call with Richard Wheeler.