All articles
Interview skills5 min read

The AI Interview Paradox: Why 92% of Candidates Rate AI Interviews Well While Others Walk Away

By , Founder and Interview CoachLast updated

Judy M. Wallis is a former senior recruiter, an interview coach and the founder of Interview Practice Lab.

Share on LinkedIn

If you've applied for a job recently, there's a good chance something in the process was run by AI — a voice screening call, a one-way recorded video, a chatbot doing initial triage. And if you've felt two completely different things about that experience, you're not imagining it. Bullhorn's report of 23 September 2026 and Greenhouse's report of May 2026 seem to contradict each other, and the gap between them is actually the most useful thing job seekers can learn from either one.

The good number: 92%

Bullhorn's 2026 GRID Talent Trends Report, released on 23 September 2026, surveyed nearly 2,300 professionals who had worked with a recruitment firm in the past three years. It found that 92% of candidates interviewed by an AI voice agent rated it at least on par with a live recruiter interview. Of candidates who'd had a positive AI experience, 93% said they'd work with that firm again, and 81% rated the experience positively overall. The AI candidates liked most wasn't a judge or gatekeeper — it was a matchmaker and a scheduler: 55% named job matching as AI's most valuable use in staffing, and satisfied candidates were three times more likely to report happiness with speed and responsiveness.

The bad number: 38%

The Greenhouse 2026 Candidate AI Interview Report (May 2026, 2,950 job seekers in the US, UK, Ireland, Germany and Australia) found that 38% of candidates had walked away from a hiring process because it included an AI interview, and another 12% said they would (as reported by Fortune). Greenhouse also found that 70% were never clearly told up front that AI would evaluate them.

These aren't the same thing

Read side by side, the reports aren't really disagreeing — they're describing two different products wearing the same label. The AI experience Bullhorn's respondents liked was interactive and transparent: a voice conversation, in real time, in service of getting matched to the right role faster. Greenhouse found the top reason candidates walked away was a pre-recorded video interview scored by AI with no human present (33%). That is a different experience from a live voice conversation, especially when candidates have no clear sense of how their answers will be evaluated.

In other words: candidates don't have a blanket problem with AI in hiring. They have a problem with not knowing what's happening to them. The same underlying technology reads as helpful when it's fast and communicative, and as hostile when it's silent and unaccountable.

What this means if you're prepping for an interview right now

If you get invited to an AI interview, the type matters more than the label:

  • Live or interactive AI (voice or chat-based screening): treat it like a real conversation. These tools are usually built to move you toward a human faster, not replace one — answer naturally, ask the AI direct questions about next steps if it allows for them, and don’t over-perform for a "test".
  • One-way recorded video with no visible reviewer: this is where nerves and distrust both spike, and for good reason — you're often talking to a camera with zero feedback. Prepare tighter, more structured answers than you would for a live conversation (a clear beginning, middle and end to every response), since there's no interviewer to redirect you if you ramble. It's also fair to ask the recruiter, before you start, who reviews these and what happens next — a company with a real process usually has a real answer.
  • Either way: you are allowed to ask how a stage works. These reports suggest that the format and whether candidates know how it works matter, not just whether AI is involved.

The takeaway

"AI interview" stopped being one thing a while ago. The gap between a 92% approval rating and a 38% abandonment rate isn’t a contradiction — it’s a pretty good filter for figuring out, in five minutes, whether the process you’re walking into respects your time or not.

Once you know which type of AI interview you've been invited to, practice for that type. If it's one-way with no feedback, the Practice Lab lets you rehearse structured, spoken answers in that same format, with a scored report on what to fix next.

Sources: Bullhorn's 2026 GRID Talent Trends Report; Greenhouse 2026 Candidate AI Interview Report; Fortune's report of 4 May 2026.

Written by Judy M. Wallis, founder and interview coach at Interview Practice Lab and a former senior recruiter.