> This is the markdown version of https://www.classet.ai/blog/ai-interview-negative-reviews > Learn more at https://www.classet.ai ![AI Interview Negative Reviews: What the Data Actually Shows](/_next/image?url=%2Fimages%2Fblog%2Fai-interview-negative-reviews.png&w=3840&q=75) # We Read Every Negative AI Interview Review. Most Had Nothing to Do With AI. We read all 1,072 written comments on negative AI interview ratings. Most had nothing to do with AI: echoey speakerphones, dropped calls, and dead mics dominate. Here's the full breakdown. [![Paul Jones](/_next/image?url=https%3A%2F%2Fassets.basehub.com%2Fe0b5701f%2F6599306507912123f90f150a8bfaaf6c%2Fscreenshot-2026-01-28-at-10.53.16-am.png%3Fwidth%3D100%26height%3D100%26quality%3D100&w=96&q=75) Paul JonesHead of Growth at Classet ](/blog/authors/paul-jones) April 8, 2026 AI Recruiting **Quick answer:** Classet has conducted hundreds of thousands of AI phone screens. We analyzed every negative rating in 10,167 post-interview responses collected January through August 2026, and read all 1,072 written negative comments. The biggest root cause of negative reviews was call infrastructure: audio problems, dropped connections, and dead microphones. Removing only the ratings where candidates described a dropped call or their own device failing, the positive rate climbs from 87.5% to 90%. Anti-AI sentiment was marginal: 5.8% of negative comments (0.6% of all candidates), and just 13 candidates, 0.13%, explicitly asked for a human. This is Part 2 of a two-part series on [AI interview candidate experience](/blog/ai-interview-candidate-experience). In [Part 1](/blog/ai-interview-candidate-experience), we published the headline number: 88% of candidates rated their AI phone screen positively across 10,167 post-interview ratings. Classet has conducted hundreds of thousands of AI interviews, and we have reviewed the candidate feedback closely. The most common unprompted comment was that it felt like talking to a real person. But we promised to come back for the other 12.5%: the 1,270 negative ratings. The ones that LinkedIn commenters would screenshot and say "See? Candidates hate this." So we did what nobody in the AI hiring debate seems to do. We read every single one of the 1,072 written negative comments. Categorized every root cause. And the story those reviews tell is not the one most people expect. ## The rating breakdown, in full Here is the complete distribution from continuous post-interview feedback collected January 17 through August 25, 2026: 05 — Rating Distribution Complete Post-Interview Feedback Breakdown Continuous feedback, January–August 2026: 10,167 live post-interview ratings Excellent 63.9% Good 23.7% Fair 4.6% Poor 7.9% 87.5% Excellent + Good 12.5% Fair + Poor 7.9% rated the experience Poor. Another 4.6% rated it Fair. That is one in eight respondents in the negative range. Those numbers are real. We are not going to minimize them or explain them away with a hand wave. But when you pull those responses apart and focus on the ones where candidates actually told us what went wrong, the reasons behind the ratings tell a very different story than "candidates hate AI interviews." ## The #1 root cause: call infrastructure, not AI We classified all 1,072 negative ratings where candidates left written feedback explaining what went wrong. Here is what they actually said: **Call infrastructure failures: 54% of negative feedback with comments.** This was the dominant theme. Audio problems led the list at 27.5%: echo on speakerphone, the AI picking up its own voice, microphones that did not transmit. Dropped and crashed calls were another 19%. And 8% of candidates explicitly blamed their own device, signal, or environment. These are failures where any phone call, human or AI, would have gone badly. **AI product issues: 23%.** Joy interrupting before the candidate finished answering, transcription errors, repeated questions, and a small number of wrong-language calls. This is legitimate product feedback and the kind of signal that actually helps improve the experience. It is also, notably, the only slice of the negative feedback that is actually about the AI's behavior, and it works out to about 2.5% of all 10,167 ratings. **The redo signal.** A recurring pattern inside the disconnection comments: candidates whose call dropped asking to interview again: "disconnected two times and didn't hold my spot," wrote one, frustrated at having to start over, not at the format. They did not walk away. They wanted another chance. That is an engagement signal. 06 — Root Causes Root Causes of Poor / Fair Ratings Live data: all 1,072 negative ratings with written feedback, Jan–Aug 2026. Values are % of negative comments. Audio / Mic Problems 28 AI Product Issue 23 Disconnection / Call Crash 19 Other / Ambiguous 16 Candidate Environment 8 Prefers a Human 6 54% Call infrastructure, not the AI 0.13% Of all 10,167 candidates prefer a human **Language mismatch: about 2%.** The interview was configured in the wrong language for the candidate. That is a setup issue, not a format rejection. As we showed in [Part 1](/blog/ai-interview-candidate-experience), when language is configured correctly, candidates respond positively: more than 140 Spanish-language candidates gave organic positive feedback without any bilingual prompting. **Preferring a human: 5.8% of negative comments.** Sixty-two candidates said some version of "I'd rather talk to a person." Across the full dataset, 13 candidates (0.13% of all 10,167) said it explicitly and emphatically. The anti-AI backlash that dominates LinkedIn threads is nearly invisible in live candidate behavior. ## The adjusted positive rate: 90% A significant share of negative ratings came from candidates who explicitly described a dropped call or their own device failing in their written comments. Their connection failed them before the interview had a chance to work. When you remove just those 289 responses (dropped calls and self-described device or environment failures, cases where any phone-based interaction would have gone the same way) the positive rate moves from 87.5% to 90%. Raw vs. Adjusted Raw vs. Adjusted Positive Rate Removing self-described dropped-call and device-failure ratings (N=289) from 10,167 live interviews Raw Positive Rate 87.5% Excellent or Good across 10,167 unfiltered ratings, including all device failures, dropped calls, and ambient noise issues. Adjusted Positive Rate 90% After removing 289 ratings where the candidate described a dropped call or their own device failing. Phone died. Bad signal. Called from a parking garage. We are transparent about this adjustment because the methodology matters. We did not remove every negative response. We kept all 295 audio-problem ratings in the negative column even though many trace to speakerphone echo rather than the AI. We only removed the ones where the candidate themselves described a dropped call or their own device as the reason for the poor experience. That is a conservative adjustment. The real number of device-related failures is probably higher, because not every candidate who had a mic problem wrote a comment explaining it. ## What hiring leaders get wrong about negative AI interview reviews There is a pattern in how hiring leaders process AI screening objections. It goes like this: Someone shares a negative data point. "12.5% rated it negatively." The room assumes that means 12.5% of candidates were offended by talking to a machine. The conversation shifts to "candidates are not ready for this." And the evaluation stalls. But that is not what 12.5% means. The data shows that the most common negative experience had nothing to do with AI at all. It had to do with the same problems that plague every phone-based interaction: bad signal, broken microphones, and noisy environments. If a candidate calls a human recruiter from a bus with a dead mic, that phone screen goes badly too. Nobody blames the recruiter. But when the same thing happens with an [AI phone screen](/product/joy-ai-powered-ats), it becomes evidence that AI interviewing does not work. That framing error is costing companies time. Teams that delay adopting [automated phone screening](/blog/automated-phone-screening) because of a perceived candidate experience risk are often comparing AI to an idealized version of human screening that does not exist. A fairer comparison: AI screening with occasional device issues versus human screening with inconsistent availability, scheduling delays, and [interviewer bias](/blog/hear-them-out-ai-phone-screenings-edge-in-reducing-hiring-bias). ## The "wanted redo" signal nobody talks about A consistent slice of the negative cohort asked to redo their interview after a technical issue. That is a small share, but the behavioral signal is strong. A candidate who hates the format walks away. A candidate who asks for another shot is telling you they were engaged enough to want a fair attempt. They were rejecting the dropped call, not the format. This pattern showed up in our [analysis of 70,000 voice AI interviews](/blog/voice-ai-in-recruiting-insights-from-70000-job-interviews) too. Candidates who experience a technical issue and get reconnected tend to complete the screen at rates comparable to candidates who had a clean first attempt. The willingness to re-engage shows up consistently. ## Pacing and environment: the feedback we already acted on The most common AI-specific complaint, inside the 23% product-issue slice, was Joy moving to the next question too quickly or talking over a candidate mid-answer. That was the most actionable feedback in the dataset because it pointed to something the product could directly address. So we addressed it. Joy now includes pacing controls that employers can configure per role. Customers can adjust how long Joy waits after a candidate finishes speaking before moving on, so the conversation matches the pace of the person on the other end. Organizations also have full control over the SMS message that goes out when a candidate first applies. That means employers can guide candidates to take the call when they have a good signal and a quiet space. You cannot fix a candidate's broken AirPods, but you can set expectations upfront so candidates know to find a good environment before the interview starts. That alone targets the single biggest source of negative feedback in the entire dataset. And the underlying voice models keep getting better. We upgrade Joy's conversation engine on a quarterly basis, moving to the latest voice AI models as they become available. Each generation improves natural turn-taking, handles interruptions more gracefully, and adapts better to different speech patterns. The pacing feedback from this study was a useful signal, but the product has already moved past it. For hiring leaders evaluating [AI recruiting tools](/blog/ai-recruitment-tools-that-actually-work-for-hourly-hiring), the question to ask is not "do any candidates have a bad experience?" Some always will, with any format. The question is whether the negative feedback points to fixable problems or fundamental flaws. Pacing and environment issues are fixable, and we have already fixed them. A systematic rejection of the format would be a different story, but that is not what this data shows. 07 — Objection Rebuttals Data-Backed Responses to Common Objections Both datasets cited where relevant Objection “Candidates won’t want to interview with an AI.” The Data 88% positive 10,167 live, unfiltered post-interview ratings (Jan–Aug 2026): 88% positive, 64% Excellent. Candidates flagging their first-ever AI interview rated it positively 96% of the time (142 of 148). Objection “Candidates will feel dehumanized.” The Data Dozens of candidates independently described the experience as feeling like a real person, unprompted — the most common theme across 4,724 written comments. Only 13 of 10,167 candidates asked for a human instead. Objection “This adds stress for candidates who are already nervous.” The Data 71% comfort at 4 or 5 / 5 In our separate structured survey (N=150, five-point comfort scale), 71% rated comfort at 4 or 5. Multiple candidates in the live feedback reported the AI format helped with nerves, not the opposite: “It definitely helps with the nerves!” Objection “What about the 12.5% negative ratings?” The Data 54% = call infrastructure Audio problems, dropped calls, and device failures dominate the written negative comments. Just 5.8% prefer a human. Many asked to redo a dropped call — engagement, not rejection. Objection “Our candidates are blue-collar — they won’t be comfortable with AI.” The Data The live dataset spans staffing, skilled trades, logistics, healthcare, and manufacturing. That’s your audience. 88% rated the experience positively. Objection “We tried AI screening and candidates complained.” The Data Complaints about AI screening most commonly trace to silence, slow process, or unclear next steps — not the AI itself. ## What this means if you are evaluating AI phone screening **Separate device problems from product problems.** When you review candidate feedback on any phone-based screening tool, ask whether the complaint is about the interaction or the infrastructure. A negative rating from a candidate with a broken microphone tells you nothing about whether candidates accept AI interviews. **The adjusted number is the one that matters for product decisions.** 90% positive is the rate that reflects how candidates experience the actual AI interaction, once you remove the cases where the interaction could not happen at all. Use that number when comparing to your current phone screen satisfaction rates, if you measure them at all. **Negative feedback that asks for a redo is positive signal.** Candidates who want to try again are candidates who bought into the format. Build your process to accommodate reconnection and you convert a dropped call from a failed screen into a completed one. **Pacing feedback is a feature roadmap.** A small percentage of candidates wanted more time to answer. That is product feedback worth acting on, and [Joy already has](/product/joy-ai-powered-ats), with configurable pacing controls, pre-interview environment prompts, and quarterly voice model upgrades. **Stop comparing AI screening to perfect human screening.** The right comparison is AI screening at 90% adjusted positive versus your current process, including the candidates who never get a call back, the ones who wait five days for a screen, and the ones who drop out because the recruiter was double-booked. Our [research on how candidates feel about AI interviews](/blog/4-findings-that-reveal-how-candidates-feel-about-ai-interviews) consistently shows that speed and consistency matter more to candidates than whether the voice on the other end is human. 09 — Best Practices Recommendations for TA Leaders Drawn from structured survey data and live feedback patterns 01 Lead with Speed 40% of abandonment = never hearing back. Set trigger timing to fire within minutes of application submission. Candidates read speed as organizational competence. 02 Set Expectations Before the Call Candidates who know they’re speaking to an AI adapt quickly. A simple pre-call SMS — “Joy, our AI screening assistant, will call you shortly” — shortens the adjustment window and improves completion rates. 03 Surface Pay and Role Details Early 26% of abandonment traces to unclear role expectations or missing pay transparency. Open every call with compensation range, schedule, and physical requirements. 04 Keep a Human in the Loop 71% are comfortable with AI screening. 12% express some discomfort. Ensure a clear, fast path from screening completion to a human recruiter — especially for candidates who ask to speak with someone. 05 Treat Candidate Experience as an Operating Metric 46.5% of candidates leave written feedback — that’s a signal in itself. Track ratings by organization, by job type, and over time. Pacing complaints are a product signal. Language-mismatch cases are a configuration issue. Separate them. ## AI interview negative reviews: what candidates actually report What do candidates complain about in AI interviews? Classet has conducted hundreds of thousands of AI phone screens across every major frontline industry. Based on 1,072 written negative comments, the #1 root cause of negative AI interview ratings was call infrastructure (audio problems (27.5%), dropped calls (19%), and candidates' own device or environment (8%)) 54% combined. AI-specific complaints like interruptions and transcription errors were 23% of negative comments, about 2.5% of all ratings, and have been addressed with configurable pacing controls and quarterly voice model upgrades. What percentage of candidates rate AI interviews negatively? Across 10,167 post-interview ratings collected January through August 2026, 7.9% rated the experience Poor and 4.6% rated it Fair, for 12.5% negative in total. When you remove ratings where candidates explicitly described a dropped call or their own device failing, the adjusted positive rate climbs from 87.5% to 90%. The negative rate actually attributable to the AI interaction itself is about 3% of all ratings. Are negative AI interview reviews about the AI or about technical issues? Our root cause analysis of all 1,072 written negative comments found 54% described call infrastructure problems (audio, dropped connections, device failures) not problems with the AI conversation itself. 23% cited the AI's behavior (interruptions, transcription errors), and just 5.8% said they'd prefer a human. Candidates who hit technical issues frequently asked to redo their interview, indicating engagement rather than rejection. How does AI phone screening satisfaction compare to human phone screens? AI phone screening achieved a 90% adjusted positive rate in our study (87.5% raw, across 10,167 unfiltered ratings). Most organizations do not systematically measure candidate satisfaction with human phone screens, which makes direct comparison difficult. However, candidates consistently report that the speed, availability, and consistency of AI screening outweigh the social familiarity of a human call, particularly for candidates who experience interview anxiety. Should negative AI interview feedback stop you from adopting AI screening? No. The data shows that most negative feedback stems from call infrastructure failures that would affect any phone-based interaction, not from candidates rejecting AI as a format. Just 0.13% of all 10,167 candidates explicitly asked for a human interviewer. The candidates who reported AI-specific pacing issues prompted product updates that are already live, including configurable pacing controls and pre-interview environment prompts. * * * _This post is Part 2 of a two-part series on AI interview candidate experience. [Part 1](/blog/ai-interview-candidate-experience) covers the 88% positive rate, first-timer reactions, stress reduction, and multilingual reach._ [![Paul Jones](/_next/image?url=https%3A%2F%2Fassets.basehub.com%2Fe0b5701f%2F6599306507912123f90f150a8bfaaf6c%2Fscreenshot-2026-01-28-at-10.53.16-am.png%3Fwidth%3D100%26height%3D100%26quality%3D100&w=128&q=75) Paul Jones Head of Growth at Classet Paul comes from an operator background running an Alpine-owned company, and brings firsthand experience with the hiring challenges Classet was built to solve. He's driven by a belief that the right technology can make meaningful work more accessible. ](/blog/authors/paul-jones) Follow Classet in Google Add classet.ai as a preferred source and our hiring research is more likely to show up in your Top Stories and AI Overviews. [Add as preferred source →](https://www.google.com/preferences/source?q=classet.ai) ## Explore More ### Use Cases - [RPO / BPO Recruiting](/use-cases/call-centers-bpo) - [Data Centers Recruiting](/use-cases/data-centers) - [Healthcare Recruiting](/use-cases/healthcare) ### Integrations - [Greenhouse](/integrations/greenhouse) - [Bullhorn](/integrations/bullhorn) - [Lever](/integrations/lever)