A practice once reserved for high-volume hourly hiring — letting a chatbot conduct the first interview — has quietly moved up-market into white-collar, full-time roles, and named employers including Coinbase, Zapier, and ManpowerGroup’s Experis are already past the pilot phase.
When Bijo Thomas, a 45-year-old tech professional from Austin, logged on for an early-stage interview this spring, he was greeted not by a recruiter but by Sophie — a realistic AI avatar with brown hair, a white top, and the ability to smile, gesture, and ask follow-up questions. Thomas described the 30-minute session as “very realistic,” according to Business Insider, which reported the story. He passed, completed two follow-up rounds with humans, and started his senior AI solutions architect role at Experis in May. Sophie, it turns out, was built in-house by ManpowerGroup.
That detail matters: this isn’t just third-party SaaS bolted onto an ATS. Large staffing and technology firms are investing engineering resources to build proprietary AI interview infrastructure — a signal that the technology is being treated as a strategic differentiator, not a commodity add-on.
Who’s Affected?
The short answer is: most active job seekers, and increasingly those pursuing professional roles, not just shift work. An April survey of nearly 3,000 active job seekers, commissioned by hiring software company Greenhouse, found that 63% of US respondents had been interviewed by AI at least once in the prior 12 months. The figure drops as you move outside the US — 57% in Germany, 54% in Australia, 47% in the UK, 36% in Ireland — suggesting American employers are leading adoption. Coinbase’s chief people officer L.J. Brock put the scale problem in plain terms: the crypto platform receives roughly 1.5 million job applications per year. “No matter how big my recruiting team is, no matter how hard we try, we can’t get to 1.5 million people,” Brock said, according to Business Insider. Coinbase began rolling out an AI interviewer called Milo in August 2025 for most roles below director level; more than 240 hires have already been sourced through that pipeline.
Zapier’s experience illustrates a different pressure point. After some job postings drew thousands of applicants within hours, the company’s global head of talent, Tracy St.Dic, said Zapier turned to AI interviews and has since screened up to five times as many candidates as its standard process allowed. Perhaps more striking: about a third of the candidates who ultimately advance to a human hiring manager would not have made the cut based on résumé and application alone. St.Dic calls these people “hidden gems” — a framing that reframes the AI layer not as a gatekeeper but as a wider funnel. For developers and engineers accustomed to being recruited aggressively, that shift in volume dynamics is worth noting. AI is already changing hiring faster than most companies can adapt their processes, and the interview stack is the latest frontier.
What Comes Next?
The vendor landscape is fragmenting fast. HR tech analyst Kyle Lagunas, cited in the Business Insider report, estimates at least three dozen vendors are now selling AI interview technology. The field includes startups like Ribbon and CodeSignal alongside established players like HireVue, which has operated in this space for years. Products range from realistic animated avatars to disembodied voice-only agents, but the value proposition is consistent: screen more applicants faster, and generate a structured output — often a numerical score plus a written summary — for recruiters to act on. Recruiters who spoke to Business Insider said the AI output functions as a guide; humans retain the final decision on advancement. Some said they also pull video or audio recordings when they want more context than the summary provides.
The fairness debate hasn’t been resolved — it’s only getting louder. The same Greenhouse survey found that 38% of job seekers had abandoned a hiring process at least once specifically because it included an AI-led interview, while another 12% said they’d do the same if required to participate. At the same time, candidates aren’t uniformly more trusting of human interviewers: 36% of US respondents said they felt evaluated differently because of age in both AI-led and human interviews, and 27% said the same about race or ethnicity. Critics, including Cornell human resources professor JR Keller, flag a different concern — opacity. “Candidates in general don’t feel like anybody cares about them as individuals,” Keller told Business Insider. Lagunas framed the stakes plainly: “We’ve got to get it right.”
There’s a compounding dynamic at play here that the individual company announcements don’t quite capture on their own. AI tools that make applying easier — the same technology powering mass-application assistants used by job seekers — have triggered a defensive arms race on the employer side. Companies facing résumé floods they literally cannot process are reaching for AI interviewers as a countermeasure. The result is a hiring loop where AI generates the applications and AI conducts the first screen, with humans entering the process only once both filters have run. This structurally depresses the signal value of the résumé and the cover letter and concentrates evaluative power in whatever the AI interviewer is optimized to measure — which, as critics note, isn’t always clearly defined or independently audited. The broader pattern of AI reshaping employment conditions is well documented; the interview layer is a newer, less-scrutinized node in that same transformation.
AI interviews are scaling hiring — but transparency is lagging
AI interview platforms differ from traditional screening tools in both scale and transparency. Avatar-based interviewers such as Sophie or Milo can operate continuously and process large numbers of candidates, producing numerical scores and written summaries with minimal human involvement. That scale, however, comes with concerns around opaque scoring, limited candidate visibility, and possible bias linked to speech, expression, or communication style. Older asynchronous video interviews offered similar scheduling flexibility, but typically relied more on recruiter review, especially after tools such as HireVue moved away from facial-expression analysis following regulatory and civil-liberties scrutiny. Human phone screens remain less scalable but more transparent, giving candidates a chance to ask questions and adjust their answers in real time, while résumé and ATS filters offer maximum automation but often penalize non-standard career paths through rigid keyword matching.
For developers, this shift is changing the tactical advice around job interviews. Alan Stein, CEO of a career-coaching firm and a former hiring manager at Google, Meta, and American Express, told Business Insider that AI-led early-stage interviews have moved from an occasional client concern to something he now hears about several times a month. Traditional interview strategies — reading body language, building rapport through small talk, and adjusting answers based on a recruiter’s reactions — become far less useful when the interviewer is a bot. Stein argues that employers may also be using AI interviews as an indirect test of AI fluency: candidates who remain professional, adapt their communication style, and engage effectively with a non-human interviewer may signal the kind of technological comfort increasingly expected in technical roles. His warning is blunt but relevant: candidates who refuse to “play the AI game” may find themselves excluded from parts of the modern hiring funnel.
The regulatory framework is still catching up. HireVue, one of the best-known companies in AI hiring, has publicly removed facial-expression analysis from its products after criticism from civil-liberties groups and regulators. New York City’s Local Law 144 now requires bias audits for certain automated employment decision tools, and similar rules may influence future legislation elsewhere. But enforcement remains uneven, and most candidates still receive limited information about how automated hiring systems evaluate them. Until disclosure and audit requirements become stronger, the lack of transparency around AI interview scoring is likely to remain one of the biggest concerns for job seekers and regulators alike.
How Serious Players Should Respond
For engineering and talent leaders building or evaluating AI interview pipelines, the Coinbase and Zapier numbers are meaningful data points — not just PR. A 5x increase in candidate throughput and a material percentage of hires who would have been screened out at the résumé stage are outcomes that justify serious architectural consideration. But deploying these tools without a documented scoring rubric, an independent bias audit, and a clear human-override mechanism is an institutional risk, not just an ethical one. Regulatory frameworks are moving, even if slowly, and the companies that get caught flat-footed will be those that treated AI interviewers as a black box procurement decision rather than an auditable system.
HR technology decision-makers should also pressure vendors on disclosure. The gap between what AI interview platforms claim to measure and what they actually optimize for is wide, and the Greenhouse survey data suggests a non-trivial share of strong candidates are already self-selecting out of AI-gated pipelines. Institutions that want top technical talent — precisely the cohort most likely to be skeptical of opaque automated systems — have a competitive reason, not just a compliance reason, to be transparent about when and how AI is used in screening. Goldman Sachs has already quantified AI’s displacement effect on the workforce; the downstream hiring architecture is now being reshaped by the same forces.
For individual candidates, the posture shift is straightforward: treat the AI interview as the interview, not as a box-checking formality. Prepare specific, structured answers — the kind that score well on rubric-driven evaluation — rather than relying on interpersonal warmth. And if you’re a developer, demonstrating that you can navigate, and even critique, AI systems thoughtfully is now part of the job pitch, starting from the first screen.
