AI Interviewers Fix Speed, Not Signal

With 63% of job seekers having interacted with an AI interviewer, they have gone mainstream in high-volume hiring, and while they do process the candidate pile fast, a fast conversation is not a validated measure of skill. What’s more, candidates are wary of being judged by AI interviewers. Our take? Let AI interviewers handle speed. Let skills assessments support the hiring decision.

AI interview adoption has climbed 13 points in six months, according to Greenhouse's 2026 Candidate AI Interview Report. For high-volume teams, the appeal makes sense. Consider that an AI interviewer never sleeps, never double-books, and can talk to every applicant on the day they apply.

"Let AI interviewers handle speed. Let skills assessments support the hiring decision. (...) Speed and signal are different problems." 

But screening every applicant quickly is not the same as knowing who to hire. We made that case earlier this year on the WRKdefined Substack when we said "Speed and signal are different problems."

Here, we take that thread further. The latest research shows AI interviewers deliver on speed and fall short on signals. Candidates are wary of them. And the format candidates say they actually want is a skills assessment.

Let’s Give Credit Where It’s Due

We want to be fair here. AI interviewers solve a real problem, and the evidence says so. The largest study to date is a natural field experiment by economists Brian Jabarian and Luca Henkel. Roughly 70,000 applicants for contact center roles in the Philippines were randomly assigned to a human recruiter or an AI voice agent. In both cases, a human recruiter reviewed the interview and made the hiring call. According to a summary from Erasmus School of Economics, AI-led interviews raised job offers by 12% and job starts by 18%. That is a meaningful result for high-volume hiring. A structured AI conversation asks every applicant the same questions, at any hour, without fatigue, and it can remove the scheduling gap where good candidates disappear.

So when it comes to the speed problem, yes, AI interviewers can solve it. The harder question is what recruiters do with the conversation once it's over.

An AI Interview Isn't a Measurement

Look closer at that same study and a telling detail appears. Recruiters rated candidates' interview performance higher when the AI ran the interview. Yet when it came time to decide, they put more weight on standardized test results, not less. Every applicant in the experiment also took a test. And the applicants who chose the AI interviewer tended to score lower on that test. In other words, even in the most favorable research on AI interviewers, recruiters still leaned on an assessment to make the call. That tracks with how these tools work. An AI interviewer records what a candidate says about their skills and may leverage pattern matching, but it cannot verify or directly test whether the candidate actually possesses those skills on its own. And certainly not against criteria proven to predict performance in the role. That limitation is made worse by the fact that 7 out of 10 people dangerously overestimate or underestimate their own skills.

The gap between saying and doing is getting wider. In Greenhouse's 2025 AI in Hiring Report, 65% of U.S. hiring managers said they had caught applicants using AI deceptively. That included reading from AI-generated scripts (32%) and showing up as deepfakes (18%). Gartner found that 6% of candidates admitted to interview fraud, and predicts that one in four candidate profiles worldwide will be fake by 2028. Greenhouse CEO Daniel Chait summed up the trend bluntly "more applications, less signal, and less transparency."

For frontline roles, this matters even more. In a contact center, retail floor, or bank branch, communication is the job. Hiring teams need to know how someone handles a frustrated customer, not how they describe what they would do. An assessment can put candidates in realistic scenarios, score open-ended responses against consistent criteria, and be validated against real on-the-job performance outcomes.

Format makes a big difference when it comes to assessing candidates. Open-ended responses scored by HiringBranch show four times the reliability as multiple-choice formats from competitors. Reliability is the key when it comes to signal, because otherwise a faster process just delivers noise faster.

Candidates Don't Trust AI Interviews

Adoption has outpaced candidate confidence by a wide margin. Greenhouse's May 2026 report found that 38% of U.S. job seekers have actually walked away from a hiring process because it included an AI interview, and another 12% say they would. The top triggers are as follows:

• Pre-recorded video interviews scored by AI with no human present (33%)
• Unclear disclosure of how AI would be used (27%)
• AI monitoring during the process (26%) 

The experience after the interview doesn’t help candidate confidence either. Of candidates who completed an AI interview, 51% never heard back. 

“Nearly half want to know a human will review their responses, and 40% want to know how they will be evaluated.” 

Only 21% believe most employers are using AI responsibly. Trust in AI as a judge is low across the board:

• Only 26% of candidates trust AI to evaluate them fairly, according to Gartner's survey of 2,918 job candidates
• 25% say they trust an employer less when it uses AI to evaluate them
• 46% of U.S. candidates want the option to request a human interview instead

To be clear, most candidates are not refusing AI interviews outright. Greenhouse's October 2026 candidate report found only 7% would refuse one, but that openness comes with conditions. Nearly half want to know a human will review their responses, and 40% want to know how they will be evaluated. 

Candidates Want an Opportunity to Show Their Skills

When candidates are asked about skills assessments, the tone changes completely. In Indeed’s 2026 Hiring Confidence Index, 54% of job seekers in India said skills assessments are a fairer way to be evaluated than a resume. Only 15% preferred the resume. Among Gen Z, nearly two-thirds favoured skills tests, work samples, or task-based assessments. Our own internal data showed that 64% of candidates preferred to be evaluated by a HiringBranch skills assessment than submitting a resume. US job seekers lean the same way. 87% told Express Employment Professionals and the Harris Poll that employers should prioritize skills over degrees.

Decades of I-O research explain why. A meta-analysis of applicant reactions found that the more job-relevant an assessment looks, the fairer candidates judge the whole selection process. That effect held across gender, age, and ethnicity. Work samples and high-fidelity simulations consistently come out on top.

What This Means for Your Hiring Stack

AI interviewers and skills assessments are not competing for the same job. Give each the work it does best.

Use AI interviewers for speed 

First contact, availability, logistics, and a consistent set of screening questions. Tell candidates upfront that AI is involved and that a person reviews the results.

Use skills assessments for the decision 

Before you put weight on any assessment, check that it:

1. Measures demonstrated skill through realistic, open-ended scenarios, not multiple-choice
2. Is validated against real on-the-job performance outcomes in roles like yours and backed by IO science
3. Shows its reliability numbers and explains how each score was reached
4. Has an independent bias audit, such as one under NYC Local Law 144
5. Keeps a human accountable for the final hiring decision

When the decision rests on validated skills data, results follow. A Fortune 50 retailer cut bad hires by 58% with HiringBranch's interpersonal, cognitive, and communication assessment. Bell Canada lowered attrition to 10% by assessing at the top of the funnel.

Speed gets candidates to your door. Signal tells you who should walk through it. High-volume hiring needs both, but only one of them should make the call.

‍

Image Credits

Feature image: Unsplash/Austin Chan