Last updated: October 5, 2026
TL;DR An AI interviewer earns its place when it conducts the first round itself and hands a human the evidence.
- An AI interviewer holds a two-way conversation and asks follow-up questions, which a recorded one-way video cannot do.
- The interview can happen over a real phone call or on video, in international and regional languages like Español or Hindi.
- For technical roles, look for coding and online assessments alongside the interview.
- Insist that every score comes with reasoning a recruiter reviews before any decision is made.
- Cheating detection is non-negotiable. Our study of 19,368 AI interviews flagged 38.5% of candidates for cheating behavior.
You have more applicants than interview slots, and the first round is where the calendar breaks. An AI interviewer is software that conducts the first-round interview itself: it asks questions, listens, follows up, and gives your recruiter a scored report to review.
We built Fabric to run a candidate's first round (resume screening, eligibility checks, and a live AI-led interview) before your recruiter or panel gets involved. It does not make the hiring decision. It gives your team a shortlist with evidence, so human time goes to final rounds.
Most buyers weigh two things: whether an AI interviewer holds a real conversation or only records answers, and whether they can trust the score it gives.
This AI interviewer buyer's guide gives both a plain answer, backed by our own interview data and three customers who shared their numbers. Start with whichever question matters most to you.
What brought you here today?
Pick the one closest to you and we will tell you where to start reading.
If you hire engineers
Our first round for technical roles includes live pair programming plus coding and online assessments. Start with the section on coding and online assessments.
If you hire at volume
First-round interviews run over a real phone call or on video, in international and regional languages. Start with the section on phone-call and video interviews.
If you need to trust the score
Each score comes with the reasoning behind it, and a recruiter reviews the report before any decision is made. Start with the section on scoring.
If you worry about cheating
Our study of 19,368 AI interviews flagged 38.5% of candidates for cheating behavior. Start with the section on cheating and AI assistance.
If you need proof it is faster
Folens went from job post to offer in 3 weeks, Furlenco hired 30 engineers in 2 months, and Skydo fills 15 positions a month with one recruiter. Start with the section on time-to-hire.
Jump to that section from the list below, or browse the whole page:
- What is an AI interviewer, and does it actually conduct the interview?
- Can an AI interview run over a real phone call or video in multiple languages?
- How are coding and online assessments handled for technical roles?
- How is each interview scored, and who makes the hiring decision?
- Does an AI interviewer catch cheating and AI assistance?
- How does an AI interviewer reduce time-to-hire?
- Where does AI interviewing fit in your hiring process, and where does it not?
- FAQ
What is an AI interviewer, and does it actually conduct the interview?
An AI interviewer is software that conducts a two-way job interview with a candidate by voice or video, then gives a recruiter a scored report. Yes, it conducts the interview itself. It asks a question, listens to the answer, and asks a follow-up based on what the candidate said.
The follow-up is what separates it from older tools. A one-way video interview, also called an asynchronous interview, records a candidate answering fixed questions with nobody on the other side. A scheduling tool only books time with a human interviewer. Neither one probes a vague answer.
At Fabric, our AI interviewer runs conversational Round 1 interviews across technical and non-technical roles, and it follows up on each answer.
For the full explainer, read our guide to what an AI interview is and how the format works. The rest of this page is about what to check before you buy one.
Can an AI interview run over a real phone call or video in multiple languages?
Yes. At Fabric, our AI interviewer runs first-round interviews over a real phone call as well as on video, and it supports both international and regional languages. A phone call needs no laptop, app or video link, so a candidate can be interviewed wherever they can take a call.
An AI voice interview is a spoken conversation between the candidate and the software, with no human on the line. When you automate first-round screening, the phone removes a step. Nobody has to install anything or find a quiet room with a webcam.
Video suits roles where you want to see the candidate. Whichever you choose, the interview is still a two-way conversation, and a recruiter still reviews the result.
Getting candidates to the interview is part of the same flow. We source candidates, reach out by email, WhatsApp and calls, and automate Round 1 interviews, so your recruiter is not chasing slots by hand.
Bulk hiring and campus hiring teams are usually the first to feel the difference. When cheating is widespread and thousands of people apply for a role, automated first-round interviews mean every eligible candidate is heard, not only the ones a recruiter had time to call.
How are coding and online assessments handled for technical roles?
For technical roles, our AI interviewer adds coding to the conversation. We offer online assessments (OAs), which are coding tests a candidate completes on their own, and live pair programming, where the candidate writes code while the interviewer asks about their choices. Both sit inside the first round.
The reason to combine them is depth. IT services and staffing teams know the pattern: the resume looks great, and the profile still gets rejected because the technical depth is missing. A coding task with follow-up questions tests whether the person can explain what they wrote.
Coding is one of several role-specific formats. Sales and customer success candidates get cold call and cold email simulations, while product and consulting candidates work through case studies and guesstimates.
For non-technical roles there are prompting exercises, which test how well a candidate works with AI tools. Technical roles also get OpenRounds, our format for testing how engineers build with AI.
Filter the table by the role you hire for to see which first-round format applies.
Which first-round format fits the role you hire for?
| Role you hire for | Our first-round format |
|---|---|
| Engineering and other technical roles | Live pair programming, coding and online assessments, and OpenRounds (AI-native skills) |
| Sales | Cold call and cold email simulations |
| Customer success | Cold call and cold email simulations |
| Product and consulting | Case studies and guesstimates |
| Other non-technical roles | Prompting exercises that test AI fluency |
For a sense of what a technical round covers, take Folens, an Irish educational publisher and one of our customers. Each candidate sat a single 50-minute AI interview that spanned coding, system design and a deep dive on experience.
If engineering is your main hiring load, our guide to the best AI interviewers for tech hiring compares the options in more detail.
How is each interview scored, and who makes the hiring decision?
After each Fabric interview, we build a report from the conversation transcript and score the candidate, with the reasoning attached to each score. We also check for 21+ cheating signals to catch unfair methods. Your recruiter reviews that report before any decision is made. We source, screen, score and shortlist. Your recruiter or hiring panel decides who moves forward.
A scorecard is the summary of how a candidate did against each criterion for the role. Explainable scoring means every score on it comes with the reason behind it. A recruiter can then check the score against what the candidate said, instead of taking it on trust.
That review step is what human in the loop means in AI candidate screening. The software does the first pass and a person approves what happens next.
Hiring teams want the scorecard, transcript and summary as soon as the interview ends. But speed only helps if they can see why a candidate scored the way they did. Otherwise, nobody knows whether the screen actually worked, and weak candidates slip through to the final round, wasting your senior interviewers' time.
So when you compare AI interviewer software, ask each vendor one thing: can a reviewer trace a score back to the candidate's own words?
Does an AI interviewer catch cheating and AI assistance?
At Fabric, our cheating detection is designed to flag 21+ signs of AI assistance during an interview and surface them to your recruiter. It's a signal for your team to weigh, and the recruiter or panel decides what to do with it. We built detection into the core of the product. It is not sold as an add-on.
The scale of the problem explains why. For our State of AI Interview Cheating in 2026 study, featured in Forbes and published in January 2026, we analysed 19,368 AI interviews conducted between July 2025 and January 2026. We flagged 38.5% of candidates for cheating behavior.
We also broke the figure down by role and by tool. Technical roles showed a 48% cheating rate against 12% for sales. Dedicated cheating assistants such as Cluely and Interview Coder made up 45% of cases, ahead of voice mode on chatbots like ChatGPT at 34%.
One finding matters more than the rest: 61% of the candidates we flagged scored above pass thresholds and would have advanced without detection.
For the mechanics, read our technical deep dive on how AI interviews detect cheating.
If you came for one figure from the study, open the question that matches it.
Looking for one figure from our study?
What does flagged mean in our study?
Which roles saw the most cheating?
Which tools were candidates using?
Would those candidates have got through anyway?
What happens when a candidate is flagged?
How does an AI interviewer reduce time-to-hire?
Time-to-hire is the number of days between a candidate entering your pipeline and accepting an offer. By our estimate, screening takes about 80% of it.
Three of our customers have shared their numbers. Folens went from job post to offer in 3 weeks for a lead full-stack engineer. Furlenco hired 30 engineers in 2 months. Skydo fills 15 positions a month with one recruiter, and closes roles in 2 weeks where it used to take 5.
The mechanism is the same in each case: first-round interviews stop waiting on a senior person's calendar.
The wider picture points the same way. SHRM research found that 85% of employers that use automation or AI say it saves them time and increases their efficiency.
Here is how each of them did it.
Which of these sounds like your team?
Pick one to see the customer whose situation matches yours.
If you have no recruiting team: Folens
Folens hired a lead engineer with no recruiter involved.
If you are hiring many engineers at once: Furlenco
Furlenco staffed five engineering stacks in parallel.
If you have one recruiter and many roles: Skydo
Skydo runs its hiring with a single recruiter.
How did Folens hire a lead engineer in 3 weeks without a recruiter?
Folens is one of Ireland's leading educational publishers. It needed a lead full-stack engineer, and its Head of Product ran every step solo, with no recruiting team and no HR support.
We screened 2,000+ applications and ran a single 50-minute cheating-proof AI interview per candidate. Ten candidates were shortlisted and the top three went on to a human interview.
The role went from job post to offer in 3 weeks, and Folens saved 50+ hours of senior time. Read how Folens did it.
How did Furlenco hire 30 engineers in 2 months?
Furlenco, one of India's largest furniture rental companies, was staffing five stacks in parallel: frontend, backend, Android, iOS and SDET (software development engineer in test). It made 30 hires in 2 months.
To get there, we screened 10,000+ candidates and ran role-specific, cheating-proof first-round interviews with custom coding questions for each stack. Senior engineers conducted none of the first-round interviews themselves.
That saved Furlenco 1,000+ senior engineering hours. Read how Furlenco hired 30 engineers in 2 months.
How does Skydo fill 15 positions a month with one recruiter?
Skydo is a fintech company working on cross-border payments in India. One recruiter was carrying 15 open roles a month.
We score applications against the job description and run a cheating-proof AI first-round interview. Skydo now closes roles in 2 weeks, down from 5, with a 40% reduction in cost-to-hire.
Love Narwar, Sales Head at Skydo, puts it this way: "In 3 months, Fabric has become an important part of our recruitment stack." Read Skydo's story.
Where does AI interviewing fit in your hiring process, and where does it not?
An AI interviewer fits at Round 1, for roles where evaluation is objective, such as engineering, sales and marketing. We are best suited to bulk hiring and campus hiring, where volume is high and interview integrity matters. For roles judged subjectively, such as design or content writing, a human interview is still better.
Can AI fully automate hiring? It can run the early stages. Our AI agents handle sourcing, outreach, resume screening, eligibility checks, scheduling and the interview itself.
The review stays with a person. A recruiter reads the report, and the recruiter or panel makes the call.
Fitting it into your stack is mostly a question about your applicant tracking system (ATS), the software where your team tracks candidates through each hiring stage. We plug into the ATS you already use, so there is no new tool to log into.
We support 20+ ATS systems, including:
- Greenhouse
- Lever
- Workday
- Ashby
- Recruitee
- BambooHR
- Bullhorn
- Ceipal
- iCIMS
- JobDiva
For your security review, we are compliant with SOC 2 and ISO 27001, two independent standards for how a company protects customer data.
Our pricing is among the most economical in the industry. When you run bulk hiring drives, you need pricing that scales with you as your requirement grows. See our pricing.
Before your next vendor call, turn this page into a short list of questions.
Questions to take to any vendor call
Tick the ones that matter to you.
FAQ
Are AI interviews a red flag?
Not on their own. What matters is whether the interview is a real, engaging two-way conversation and whether a recruiter reviews the result before any decision is made.
Are AI job interviews real?
Yes. An AI interviewer conducts an actual first-round interview by phone or video, and one of our studies alone covered 19,368 of them.
What types of roles work best with AI interviews?
Roles with objective evaluation, such as engineering, sales and marketing, work best at Round 1. For roles judged subjectively, such as design or content writing, a human interview is better.
What insights do recruiters receive after an AI interview?
At Fabric, we give your recruiter a report built from the conversation transcript, with scores and the reasoning behind each one. Any cheating flags are surfaced for your recruiter to weigh.
Does the AI interviewer integrate with our ATS?
Yes. At Fabric, we read from and write back to 20+ ATS systems, including Greenhouse, Lever, Workday, Ashby, Recruitee, BambooHR, Bullhorn, Ceipal, iCIMS and JobDiva.
What is the 80/20 rule in interviewing?
The 80/20 rule in interviewing is a rule of thumb that the candidate should do about 80% of the talking and the interviewer about 20%. It is a habit from structured interviewing that keeps the interviewer asking and listening, not a measured standard.
Related Posts
- What is an AI interview? The complete guide for 2026
- Behind the scenes: how Fabric's AI interviewer evaluates candidates
- State of AI interview cheating in 2026: insights from 19,368 interviews
- Best AI interviewers for tech hiring
- Best AI interview platforms compared
What this means for your first round
An AI interviewer is worth judging on one test first: does it hold the conversation, or does it only record one? A tool that asks follow-up questions can test an engineer's code. It can also be designed to flag signs of AI assistance for your recruiter to weigh.
The second test is what happens afterwards. If you cannot trace a score to the candidate's own words, you cannot defend it, and a first round with no integrity check is easy to game.
Whichever vendor you choose, keep your recruiter in the review seat. The first round is changing fast. The decision should stay where it is, with your team.