AI Interviewers: The New Frontier in Job Recruitment

Meet your AI interviewer: Seattle startup Humanly unveils video bot that grills job applicants

"More human than human" is the tagline for the company that produces replicants, the bioengineered humans that run amok in the classic science-fiction film "Blade Runner." "More Humanly than human" could very well be the pitch on a new product from Seattle-based recruiting software startup Humanly, which unveiled its new "AI interviewers" on Wednesday. Billed as an "AI-powered solution that allows you to interview every candidate over video, any time of day, without sacrificing quality, structure, or human insight," Humanly views its agentic AI solution as a scalable way to unburden hiring teams that are overwhelmed by intense candidate interest, tight timelines and high expectations. Humanly touted the launch in a blog post and CEO Prem Kumar talked it up on LinkedIn, sharing a video in which an AI version of Humanly People Operations Manager Sarah Bernstein interviewed the real Bernstein about a fictional role as a customer support representative at a fictional company called Acme. "Now you can interview the world, equitably, and spend human time on the best fits," Kumar wrote, citing data that says that only about 5% of job candidates get to a human job interviewer. Founded in 2018, Humanly already uses a variety of automation software to help companies screen job candidates, schedule interviews, automate initial communication, run reference checks, and more. It competes against a wide swath of recruiting startups and larger platforms and has raised $24 million to date. Many of Humanly's customers are large companies outside of the tech industry that hire in high volumes. In the 5-minute demo video (below), the AI interviewer comes across as a realistic-looking video version of Bernstein, if not a tad bit robotic. The AI stares directly into the camera, occasionally tilting its head, wrinkling its brow and offering up slight smiles as the real Bernstein answers questions and describes how she's suited for the (fake) role. There's no AI toddler wandering into the background and no AI dog barking in the distance. The video is devoid of any human-generated spontaneity or emotion such as surprise or laughter. It signs off like it's reading an email. Humanly said the AI interviewer was built using analysis from more than 4 million interview interactions and its work "with leading recruiters to define what makes the 'anatomy of a strong interview.'" The startup also partnered with Katherine Hilton, a linguistics lecturer at Stanford, and Grin Lord, co-founder and CEO of Bellevue, Wash.-based mpathic, a startup whose software analyzes workplace communication. The goal was to reduce bias and improve predictive accuracy, according to Humanly. After an AI interviewer meets with a job candidate, the program provides feedback to a recruiter, highlighting strengths, gaps, and fit. It generates a candidate score and offers a recommendation on whether the candidate should move on to the next round of interviews. In comments on his LinkedIn post, some expressed misgivings about what Kumar and Humanly have built, while others celebrated it as a breakthrough. "Unfortunately, I'm not as impressed," wrote Aisha Bower, a product designer and AI enthusiast, according to her LinkedIn profile. "It falls short of a genuine human experience and could leave interviewees feeling less than seen and heard. They may wonder if the AI agent will represent them accurately to the employer. They may also feel that a company utilizing this for screening is too impersonal, undermining trust." Bower suggested it could be a good tool marketed to people privately trying to test out their interview skills in preparation for a live session. Startups such as Seattle-based Yoodli offer AI roleplay products along these lines. Kumar responded by saying that the product is not intended as an AI vs. human comparison. Rather, it's an AI vs. being ignored scenario, in which candidates at least get to feel as if they're moving along in a job search process and not being ghosted. "The end goal for us is not to feel like a genuine human. I think that boundary needs to be clear," Kumar wrote. "But to be human enough to make the candidate feel comfortable in providing detailed enough inputs to be fairly evaluated. Sometimes even more comfortable than a convo with an actual human." Kumar also noted that there's a candidate fairness aspect at play, theorizing that if a job posting attracts 4,000 inbound applicants, not only will many be ignored, but interview bias will creep in during human-to-human interaction. The New York Times summed up the trend particularly well in a story earlier this month: "You thought artificial intelligence was coming for your job? First, it's coming for your job interviewer." While resume screening and meeting scheduling has become an accepted, automated aspect of job searches, autonomous interviewers are popping up from a variety of companies. It's adding AI to a part of the process that has long seemed to most need a human touch, as the Times put it. Some job seekers who spoke to the Times called their interactions with AI "dehumanizing," or they wondered whether there was a job at all and if they were just part of an experiment designed to train the AI. In response to one comment on LinkedIn, about whether her AI could just do all the work while she's on vacation, Bernstein joked that she's not sure if Kumar would notice the difference at this point.

Welcome to your job interview. Your interviewer is AI

When Jennifer Dunn landed an interview last month through a recruiting firm for a vice president of marketing job, she looked forward to talking to someone about the role and learning more about the potential employer. Instead, a virtual artificial intelligence recruiter named Alex sent her a text message to schedule the interview. And when Dunn got on the phone at the appointed time for the meeting, Alex was waiting to talk to her. "Are you a human?" Dunn asked. "No, I'm not a human," Alex replied. "But I'm here to make the interview process smoother." For the next 20 minutes, Dunn, a marketing professional in San Antonio, answered Alex's questions about her qualifications -- though Alex could not answer most of her questions about the job. Even though Alex had a friendly tone, the conversation "felt hollow," Dunn said. In the end, she hung up before finishing the interview. You might have thought AI was coming for your job. First it's coming for your job interviewer. Job seekers across the country are starting to encounter faceless voices and avatars backed by AI in their interviews. These autonomous interviewers are part of a wave of artificial intelligence known as "agentic AI," where AI agents are directed to act on their own to generate real-time conversations and build on responses. Some aspects of job searches -- such as screening resumes and scheduling meetings -- have become increasingly automated over time, but the interview had long seemed to be the part of the process that most needed a human touch. Now AI is encroaching upon even that domain, making the often frustrating and ego-busting task of finding a job even more impersonal. Talking to AI interviewers has "felt very dehumanizing," said Charles Whitley, a recent computer science and mathematics graduate from Santa Clara University who has had two such conversations in the past seven months. In one interview, for a software engineering job, he said, the AI voice tried to seem more human by adding "ums" and "uhs." It came across as "some horror-movie-type stuff," Whitley said. Autonomous AI interviewers started taking off last year, according to job hunters, tech companies and recruiters. The trend has partly been driven by tech startups like Ribbon AI, Talently and Apriora, which have developed robot interviewers to help employers talk to more candidates and reduce the load on human recruiters -- especially as AI tools have enabled job seekers to generate resumes and cover letters and apply to tons of openings with a few clicks. AI can personalize a job candidate's interview, said Arsham Ghahramani, the CEO and a co-founder of Ribbon AI. His company's AI interviewer, which has a customizable voice and appears on a video call as moving audio waves, asks questions specific to the role to be filled, and builds on information provided by the job seeker, he said. "It's really paradoxical, but in a lot of ways, this is a much more humanizing experience because we're asking questions that are really tailored to you," Ghahramani said. Propel Impact, a nonprofit in Vancouver, B.C., that teaches young people about financial investing, began using Ribbon AI's interviewer in January. That allowed the organization to screen 500 applicants for a fellowship program it offers, far more than the 150 applicants who were interviewed by people last year, said Cheralyn Chok, Propel Impact's executive director. "There's no way we would have been able to successfully recruit and set up offers to 300 people to join our program," she said. Chok said the AI interviews also saved applicants the hassle of doing multiple interviews with outside financial firms to determine their fellowship placements. Instead, Propel Impact sent the recorded AI interviews to those companies. And there was still a human element, she said, since the organization told applicants that they could ask her team questions at any point. Humans cannot ultimately be taken out of the hiring process, said Sam DeMase, a career expert at ZipRecruiter, an online job board. People still need to make the hiring decisions, she said, because AI may contain bias and cannot be trusted to fully evaluate a candidate's experience, skills and fitness for a job. At the same time, more people should expect AI-run interviews, DeMase said. "Organizations are trying to become more efficient and trying to scale faster, and as a result, they're looking to AI," she said. That's bad news for people like Emily Robertson-Yeingst of Centennial, Colo. In April, she was interviewed by an AI named Eve for a role as vice president of product marketing at a software company. Robertson-Yeingst was required to keep her camera on during the call, with Eve showing up as a little gray box in a corner of the screen. Eve asked Robertson-Yeingst to talk about herself, and then later asked her to "tell me about a time you had to build a team from scratch," among more than half a dozen questions. After almost an hour, Robertson-Yeingst asked Eve about next steps in the hiring process. Eve was unable to answer, she said. In the end, Robertson-Yeingst never heard back -- from a human or an AI -- about the job, which she later saw posted again on LinkedIn. The entire experience left her feeling "used," she said. "It starts to make you wonder, was I just some sort of experiment?" she said. "Were you just using me to train the AI agent? Or is there even a job?" Others said they liked talking to AI interviewers. James Gu, a college student majoring in business in Calgary, Alberta, spoke to a robot interviewer for a summer analyst position through Propel Impact in February. Being drilled with questions by someone stresses him out, he said, so part of him was relieved not to speak with a person. During the interview, the AI asked Gu to "tell me more" about his experience running an entrepreneurship club on campus. He said he had felt freer to "yap" to the AI. "It felt like it was interested in learning about me," said Gu, who landed the job. Dunn has had about nine job interviews over the past two months, she said. Only one was with an AI like Alex, she said, for which she was "grateful." Given the choice, she never wants to interview with AI again. "It isn't something that feels real to me," she said.

Would you trust your job search to a bot?

As artificial intelligence creeps into every aspect of the hiring process -- from candidate screenings to video interviews -- a new AI agent released by the startup Jobright.ai is aiming to make endless scrolling on job sites a thing of the past. The bot is part of a growing supply of AI-powered job search tools, tapping into applicants' frustrations with protracted job searches and a stagnant labor market. Here's how it works: Once a user uploads a résumé, informs the bot of the kinds of jobs they are looking for and answers some basic questions, the bot starts compiling job postings which match the applicant's criteria and will send up to 50 listings each week. Along with showing the user the responsibilities and requirements of the roles, Jobright.ai aggregates relevant information for each listing, with details on the company's culture, history, funding and employee satisfaction levels. If the user approves any of its suggestions, the bot can complete an application, including short-answer prompts, in less than a minute. Over time, the AI agent will become more attuned to a user's preferences as the person pursues or skips past different job opportunities, said Jobright.ai co-founder and Chief Executive Officer Eric Cheng. It's "similar to how you work with Tinder," he said. To assess whether an applicant is a good fit for a given job listing, Jobright.ai translates the applicant's skills and experience into a "qualification score," a percentage which appears in a ring of mint green next to each job posting. The bot has to determine that a user is at least 60% qualified for a job, in order for it to recommend the listing. Of course, what a qualified applicant looks like can vary across industries. In tech spaces, coding-language proficiency and machine-learning skills are prized. In other sectors, strong interpersonal skills might be weighted most heavily, but they're also notoriously difficult to measure. Titles and other markers of experience also can vary by industry and from company to company. A vice president at a large bank, for example, might have a very different experience profile from a vice president at a startup. "Sometimes the AI doesn't know that kind of difference," said Cheng, who enlisted career coaches and recruiters to help the company develop its qualification-scoring system. "We need to fine tune the AI models with the actual input from the expert." Jobright.ai, which is currently geared more toward tech and engineering roles, is financed in part by the venture-capital arm of Recruit Holdings Co., the Japanese company that owns the jobs board Indeed and the employee-reviews site Glassdoor. On an earnings call in May, the parent company previewed a new Indeed "career scout," which, like Jobright.ai, will customize resumes and fill out job seekers' applications for them. It hasn't yet been released. Rival LinkedIn, which introduced AI resume feedback and cover letter drafts for paying subscribers last June, rolled out a feature earlier this year that allows users to tell an AI search tool about the jobs they want, as opposed to relying on strategic keyword searches. By making better matches between prospective candidates and employers, Cheng said, AI can address the concerns not only of job seekers but of hiring managers, who are often inundated with resumes, making it difficult and time-consuming to separate serious contenders from unqualified applicants. Cheng said Jobright.ai is currently beta-testing a recruiter-focused interface and piloting it with about 30 employers, mostly small to medium-size startups. Cheng said most of the jobs listed by employers directly on Jobright.ai's site were filled within 30 days. "Instead of giving them 500 applicants, most of them aren't qualified, we only probably recommend maybe 20 to 30 candidates, similar to like how a headhunter works," Cheng said.

Got a job interview? Be ready to impress a bot

Artificial intelligence is now handling initial job interviews in many tech companies. Firms like Fractal Analytics and UST are using AI to screen candidates. AI conducts interviews, designs tests, and checks for cheating. It also analyzes candidate communication. This technology helps companies find better candidates and reduces hiring time. Recruiters can then focus on candidate fit. Imagine walking into your first job interview and the person across the table isn't human. Instead, it's a smart AI bot, quietly sizing you up before any recruiter even sees your face. This might sound like sci-fi, but it's becoming reality as companies like Fractal Analytics, UST, and Happiest Minds hand over the early rounds of hiring to artificial intelligence, TOI has reported. The tech world is bursting at the seams with applicants, and companies are struggling to find the right fit. "Nearly 50% of mid-to-large tech companies are now using AI in some part of the hiring process - mainly in screening and assessments - though many are still in early or pilot stages," TOI quoted Murali Santhanam, CHRO at Ascent HR Technologies, as saying. What started as simple resume scanning has now morphed into AI conducting full-on interviews. "Five years ago, AI was just scanning resumes. Today, it's effectively acting as an interviewer," said Savita Hortikar, HR head at FractalAI. Their AI, Ikigai, designs fresh, role-specific tests for every candidate on the spot, cracking down on cheating and boosting hiring quality. "Every candidate now gets a fresh test. It has significantly improved the quality of hiring," Hortikar told TOI. The perks? Flexibility. Candidates get flexible interview timings, including late nights when human panels are scarce. "Most candidates prefer late-night interviews. Finding a tech panel at that hour is hard," admits Rajesh Chandran Sogasu, head of talent acquisition at Happiest Minds. Their AI steps in, posing tricky scenario questions and scoring responses, leaving humans to review the top picks later. At UST, the AI does double duty, not just testing candidates but keeping a sharp eye on exam integrity. "It tells us if the answers are original or AI-generated. We've built mechanisms to detect tab-switching, eyeball movement, and impersonation," says Kishore Krishna, VP, talent acquisition. In fact, last year, UST used AI to flag over 500 fraudulent job offers, based on video audits that detected mismatches between the people interviewed and those who eventually received the offers. This tech has already slashed hiring times by 20%, with hopes to cut it in half, as per the TOI report. The AI's powers don't stop at code. Fractal's Ikigai also analyzes how candidates express themselves, probing why they want a new job and decoding their tone and mindset. "The bot might ask why the candidate is looking for a new role. It analyses tone, structure, and coherence to offer insights into the candidate's mindset," Hortikar says. So where does this leave recruiters? "AI is surfacing great candidates -- even those with poorly written resumes. The recruiter focuses on selling the job and gauging fit," says Milind Shah, MD at Randstad Digital India. (With inputs from TOI)