For many job seekers, the first conversation in the hiring process is no longer with a recruiter — it’s with an artificial intelligence system that conducts the interview and evaluates the applicant’s responses.
How to Prepare for an AI Interview
- Plan for common questions
- Keep your responses relevant and concise
- Practice speaking out loud
- Practice with a mock AI interview tool
- Avoid relying on AI-generated scripts
- Be mindful of the “feedback void”
AI-led interviews are the latest escalation in the so-called AI “doom loop” in which more AI use on one side of the hiring desk begets more AI use on the other. Candidates are increasingly relying on AI-generated applications and auto-apply tools to boost their chances of standing out in a job market where more applicants are competing for fewer roles. Meanwhile, recruiters — faced with a deluge of nearly identical AI-generated resumes and cover letters — are relying less on a candidate’s written words and more on their ability to answer questions about their skills and experiences on the fly. By inviting dozens or even hundreds of applicants to participate in AI interviews, recruiters are hoping to identify strong candidates without blowing up their calendars.
But what does it take for a job seeker to make it past this robotic gatekeeper and move on to the next round? As AI interviews become more commonplace in this competitive job market, applicants should know what to expect, how these tools work and how to showcase their skills without losing their authenticity.
The Rise of AI Interviews
AI interviews are job interviews that are administered by a conversational AI agent instead of a human recruiter or hiring manager. Developed by companies like HireVue and Jobma, these platforms ask candidates questions, evaluate their responses — sometimes including their speech patterns and facial expressions — and submit a report card to the hiring team. AI interviews typically happen early on in the hiring process as a sort of screening measure, allowing hiring teams to sort through large numbers of applicants and find those they would like to speak with themselves.
Initially more popular for hiring front-line, hourly workers, AI interviews are increasingly used to recruit knowledge workers across professions. Nearly two-thirds of the U.S. job seekers surveyed by hiring platform Greenhouse say they have participated in an AI job interview — but respondents have mixed feelings about them. About 38 percent of U.S. candidates said they have withdrawn from a hiring process because it included an AI interview, with some of their top concerns being a lack of clarity and transparency around AI’s role as a decision-maker. And yet, only 19 percent of respondents said there should be less artificial intelligence in the hiring process.
In fact, nearly 80 percent of the 70,000 applicants surveyed in a University of Chicago study preferred to be interviewed by a voice AI platform instead of a human interviewer. Researchers posited that the candidates may view these systems as more objective or predictable than a person, or that they like the convenience of scheduling an interview that same day or outside of normal business hours. One company told Wired that 24 percent of its AI interviews take place between 10 p.m. and 2 a.m., illustrating a demand for flexibility that would never be possible with a human recruiter.
While AI interviews are necessary to help hiring managers sort through mountains of applications, Deepali Vyas, global head of data and AI executive search at talent advisory firm ZRG Partners, told Built In that it can also dehumanize the process for employees seeking an authentic, human connection with a potential employer. Applicants now view the process more like a test or an audition than a conversation, she said, and feel like they need to “game the algorithm” instead of building a narrative.
That makes sense, because “at a very rudimentary level, [AI] is pattern matching between what the job description says and how much of a fit you are for that pattern,” Shama Hyder, an applied AI evangelist at Wispr Flow, told Built In. And unlike a human recruiter, the algorithm is unlikely to be charmed by a candidate’s personality or take a chance on a candidate whose background doesn’t neatly align with the job description.
“It’s a double-edged sword in that you’ve got the ability to have an objective review of your work, but the downside is you don’t really have the opportunity to explain outliers or build rapport.” Hyder said.
What to Expect in an AI Interview
If you are selected for an AI-led interview, the company may pair you with an avatar that asks you questions, or it may ask you to respond to a question that appears on the screen. It will then give you a small window of time to gather your thoughts and then respond within a specified timeframe — typically two minutes. The system may then create a scorecard evaluating your responses on a variety of factors, potentially including your facial expressions and body language. The interview is usually recorded for a recruiter or hiring manager to review afterward.
These are the three most common AI interview formats:
- Asynchronous interviews: Candidates are given a standard set of questions, and then they record their responses. Candidates typically have the option of re-recording their answer before submitting it to the system for evaluation. Asynchronous interviews were the first AI interview tools to hit the market. They’re still used today, but are increasingly being replaced with “live” or “conversational” AI interviews.
- Live video interviews: These interviews are administered in real-time by an AI, which is sometimes portrayed as a human-like avatar. Unlike asynchronous interviews, these systems can ask follow-up questions based on the applicant’s responses. The follow-up questions are not necessarily hard-hitting, though: one comedian recorded himself giving nonsensical responses, and the interviewer simply repeated the gibberish phrases back to him without questioning their meaning.
- AI-assisted interviews: Even if an applicant is talking to a human interviewer, it’s increasingly likely that the person has an AI agent working behind the scenes to transcribe or summarize the interview. While some may not consider this to be a true AI interview, it shows that learning to communicate clearly and concisely for AI systems will be more important than ever as employers incorporate more of these tools into their hiring processes.
How AI Evaluates Applicants
AI interview systems use proprietary black-box algorithms to evaluate candidates on various criteria, which can typically be customized to a company’s preferences. The platform will then produce a report card showing how each candidate scored on the selected criteria, or it may pull out snippets of the applicant’s most relevant responses. While the specific criteria and algorithms may vary, AI interview platforms generally judge applicants on the content of their responses, their vocal delivery and their body language.
Content
AI systems are primarily focused on the clarity, relevance, structure and consistency of the applicant’s responses.
- Clarity: If an AI system can’t understand what you’re saying, it could transcribe your words inaccurately and misconstrue what you’re trying to say.
- Relevance: AI interview platforms are listening for keywords and other signals that your skills and experience are aligned with the competencies of the job and the company culture.
- Structure: Large language models like a story with structure — a beginning, middle and end. This is why the STAR method can be helpful for organizing your responses.
- Consistency: AI interview platforms view consistency as a signal of trustworthiness or authenticity, so it is listening to ensure that your responses don’t contradict your other statements in the interview, on your resume or on your LinkedIn profile.
Vocal Delivery
Some AI systems also analyze the tone of an applicant’s voice. So if an applicant is speaking in a monotone voice, for example, they will not score as high as an applicant that speaks with energy and enthusiasm. It might also pay attention to how long it takes you to come up with an answer, how fast you’re talking or whether you’re using a lot of filler words.
Body Language
AI systems also analyze nonverbal communication. Eye contact, in particular, can convey trustworthiness, as these systems are particularly sensitive to detecting whether an applicant is reading from a rehearsed script or seeking real-time interview help from a separate AI platform. Some systems might also look at facial expressions, hand gestures and posture. That being said, experts advise against putting on an inauthentic performance to simply please the AI.
“You don’t need to radically change who you are to be able to get a job, but you need to be mindful of what the person or the algorithm is looking for,” Lili Foggle, director of the Interview Institute at the Professional Association of Resume Writers and Career Coaches (PARWCC), told Built In.
How to Prepare for an AI Interview
Preparation is everything in an interview, AI or otherwise. Here are some tips to make sure you’re optimizing your responses for this new format.
Plan for Common Questions
Preparing for an AI interview is similar to preparing for a regular interview: read the job description, research the organization and reflect on what you can bring to the role. Prep for common screening interview questions, as well as some common behavioral interview questions that ask for specific examples of your impact. It can be difficult to recall past experiences under pressure, so it’s always helpful to come ready with about five or six stories about your biggest accomplishments, a failure, a conflict, a difficult decision or something you built or changed, Vyas said. If you’re having trouble sorting through the many potential behavioral interview questions, start with those that are most applicable to the job description in question.
Keep Your Responses Relevant and Concise
Because you are talking to a machine, you may have to be more direct than you would with a human. The AI system cannot infer meaning the way humans do, so grounding your answers in keywords from the job description can signal that you are a good fit for the role. At the same time, Vyas advises applicants against trying to “game the algorithm” by stuffing too many keywords into their answers.
“Don’t change who you are because there’s a machine on the other side,” Vyas said. “I think one of the biggest risks is that candidates suddenly become robotic themselves.”
To deliver your responses with the most impact within the platform’s time constraints, structure your responses using the STAR method: describe the situation, your assigned task, what actions you took and the result of your actions, quantifying your impact as much as possible.
Practice Speaking Out Loud
Written responses may look good on paper, but Foggle suggests applicants practice saying their responses out loud to see if they sound natural and fall within the recommended time limit. By speaking through your answers, you can also become more skilled at talking about your professional accomplishments, which will help you answer unexpected questions with more confidence and authenticity.
“You need to practice interviewing as a speaking skill. It is not a writing skill. It is not a memorize-and-recite skill. It is a conversational skill,” Foggle said. “The only way to get better at that skill is to practice speaking.”
Practice with a Mock AI Interview Tool
One of the best ways to practice is by using a mock AI interview tool, which are available online through Teal, Final Round and other websites. These platforms can help you familiarize yourself with the idiosyncrasies of AI interviewing, like speaking to an avatar and limiting your responses to a specified time limit.
Mock AI interviewers can typically generate custom questions based on a job description, and they offer feedback that can help you hone your responses. If you don’t want to use one, Foggle suggests speaking into your phone’s voice memo app or any other transcription tool to see if AI is repeatedly misunderstanding certain words or phrases.
Avoid Relying on AI-Generated Scripts
Other AI tools like ChatGPT or Claude can also be useful for interview preparation. They can help you research the company, generate sample interview questions or even hone your interview answer. But Foggle and other experts advise against using these tools to generate a script for you, as they will be limited to generic responses that don’t reflect the unique skills and experiences that you can contribute to the organization.
Plus, as Foggle noted, “reciting a script is never going to be authentic; it’s always going to be performative.” When preparing potential interview responses, Foggle suggests jotting down bullet points that you can refer to rather than writing out and reading full sentences. Some job-seekers think authenticity doesn’t matter when talking to an AI avatar, but almost all of these platforms have implemented eye-tracking software to penalize candidates who read from a prepared script or use an AI interview assistant that feeds applicants answers in real-time during an interview.
Be Mindful of the ‘Feedback Void’
As humans, our communication is built around talking to other people. Subconsciously, we are often reading the interviewers’ facial expressions and body language in real-time to gauge whether our words are resonating. This feedback can provide clues that help us decide whether the interviewer understands or agrees with you, or if they are confused or losing interest.
Some applicants really struggle with the “feedback void” AI interviewers create, Foggle said, and they may “spiral and talk faster” as a result. For these applicants, she recommends they tape a picture of someone they have warm feelings about — but also someone they want to impress — near their computer’s camera.
“Putting a little picture near a camera reminds people to stay looking at the camera and maybe even smile occasionally, which can increase their communication scores, their authenticity scores and just in general kind of warmth,” Foggle said.
AI’s Potential for Bias
The notion of one’s professional livelihood hinging on a black-box algorithm’s evaluation of their vocal patterns and facial expressions understandably raises concerns about potential AI bias in the hiring process.
AI models may not accurately transcribe the words of somebody who is not a native English speaker, has an accent or has a speech impediment, for example. Additionally, some people on the autism spectrum naturally speak in a monotone voice, so they would be penalized by an AI interview platform that scores applicants on the pitch, intonation or enthusiasm in their voice.
The facial analysis features in some of these systems have also come under fire. HireVue, one of the most popular AI interview platforms on the market, stopped using facial analysis technology in 2021 after a public interest group filed a discrimination complaint with the Federal Trade Commission. Other platforms that attempt to predict an applicants’ qualifications before an interview, like Eightfold and Workday, have also been sued amid allegations of algorithmic discrimination.
In fact, the potential for bias has led several states to pass legislation around AI in the hiring process. Illinois, for example, requires employers to obtain an applicant’s consent to AI evaluation and to explain how the system will evaluate them. Maryland, meanwhile, requires employers to obtain an applicant’s consent before using facial analysis technology. Lawmakers in New York City, Colorado and a growing number of jurisdictions have passed legislation to regulate the use of AI in hiring, which Foggle said is “really like the Wild, Wild West.”
“[AI usage by both recruiters and job seekers] is really escalating,” Foggle said. “It’s broken, and this is why it takes six to eight months to get a job right now.”
Frequently Asked Questions
What happens in an AI interview?
Candidates may be asked questions by an AI avatar or respond to written questions. Depending on the platform, applicants may have a limited amount of time to prepare and answer. The system then evaluates their responses and may create a scorecard for recruiters to review.
What are the different types of AI interviews?
AI interviews generally fall into three formats: asynchronous interviews, in which candidates record answers to standard questions; live video interviews, in which an AI conducts a real-time conversation; and AI-assisted interviews, in which AI helps a human interviewer transcribe or summarize the conversation.
What does AI look for in an interview?
AI interview platforms can evaluate the content of an applicant’s responses, vocal delivery and body language. Depending on the system, this can include the clarity, relevance, structure and consistency of answers, as well as factors such as speaking pace, filler words, eye contact, facial expressions, hand gestures and posture.
How should I prepare for an AI interview?
To prepare for an AI interview, job seekers should practice answering common interview questions, using the STAR method to structure their responses. They should keep their answers relevant and concise to comply with the specified time limits, and they should avoid rambling as much as possible. Job seekers should also practice maintaining steady eye contact, as some AI systems penalize candidates who appear to be reading from a script.
Should I use AI to prepare for an AI interview?
AI tools can help candidates research a company, generate sample interview questions and improve responses. However, the article recommends against using AI to write a script, because generic responses may not reflect a candidate’s specific experience and rehearsed answers can sound inauthentic.
