An AI interviewer runs the first conversation with a candidate for you. It asks structured questions about the role, follows up on vague answers, and returns a transcript, scores and a recommendation. Used well, it replaces the phone screen: the 20 minutes per candidate that a small team cannot find when a role attracts two hundred applicants. It does not replace the decision, which stays with you.

The problem it solves

Posting a job is easy and reading the response is not. A remote role can draw hundreds of applications in a week. Most teams deal with that by reading the first fifty CVs and phone-screening eight people, which means the hire is chosen from whoever applied early and wrote a tidy CV.

Two steps consume the time: sorting applications and running first conversations. Both can now be handled by software, and both produce better results when the software works from criteria you set rather than from its own idea of a good candidate.

How an AI first round works

  1. You define the role. Must-have skills, years of experience, location and anything else that matters. This is the standard every applicant is measured against.
  2. Applicants are ranked. Each CV is compared with those requirements and given a score with the reasons behind it, so you can see why someone is third and not thirtieth.
  3. The interview runs when the candidate is ready. No scheduling. The candidate opens a link and talks to the interviewer in their own time.
  4. You read a report. A transcript, scores, quoted highlights and a recommendation, usually within minutes of the interview ending.

What to look for in an AI interviewer

Questions grounded in your role and the candidate’s CV

A generic question bank produces rehearsed answers. The interviewer should read the job description and the candidate’s CV and ask about the work they claim to have done, plus any questions you add yourself.

Real follow-ups

The value of a phone screen is the second question: what was your part, why that approach, what was the result. A tool that only plays a fixed list and records answers is a video form, not an interviewer.

A practical task, when the role needs one

For roles where the work can be shown, a short typed task during the interview tells you more than another ten minutes of talk.

Integrity signals you can read for yourself

Candidates now have AI tools too. A useful report shows you what happened: whether the screen share stayed on, whether text was pasted into the task, whether the window lost focus, and whether answers contradicted the CV. You should see the evidence, not just a number.

Pricing that follows usage

A monthly seat licence makes sense if you hire every week. If you hire a few times a year, look for per-interview or per-minute pricing with no subscription.

Fairness, consent and the law

This is the part to get right before you send a single invitation.

  • Keep a person in the decision. Treat the report as evidence for a human reviewer. AI output can be wrong, and it should never be the only basis for rejecting or hiring someone.
  • Tell candidates. Say in the invitation that the first round is conducted by an AI interviewer, what is recorded, and who will see the result.
  • Score the answer, not the accent. Communication scores should measure clarity and structure. Fluency in a particular variety of English is not a job requirement unless the role makes it one.
  • Offer an alternative. Some candidates will need adjustments for disability, connectivity or equipment. Have a human route available.
  • Check your local rules. Several jurisdictions regulate automated tools in hiring, and requirements on notice, consent and bias audits differ. You remain responsible for a fair and lawful process, whatever tool you use.

Setting up a good AI first round

  • Keep it to 10 or 15 minutes. Long enough for four or five substantive questions, short enough that good candidates finish it.
  • Add three questions of your own. The ones you always end up asking on the phone.
  • Decide what a strong answer looks like before you read any reports. It stops the first impressive transcript from setting the bar.
  • Interview a shortlist, not everyone. Rank first, then invite the top of the list.
  • Read the quoted highlights, not only the score. Two candidates on 78 can be very different hires.

What it costs

Pricing in this market comes in three shapes: an annual platform licence, a monthly plan with a set number of interviews, and pay-per-use. For a team hiring a handful of roles a year, pay-per-use is almost always cheaper, because you pay nothing in the months you are not hiring.

As a worked example, Hyrovo charges $0.60 per interview minute, billed on the minutes actually used, after two free interviews. A 10-minute first round costs up to $6. Screening a shortlist of 15 candidates costs up to $90, and posting the job and ranking every applicant is free. Compare that with the hours a hiring manager would spend on 15 phone screens.

How Hyrovo does it

On Hyrovo you post a job free, and every applicant is ranked against your requirements with a breakdown across skills, experience, domain and evidence. You can then ask candidates to complete an interview with Dami, our AI interviewer, as a text chat or a video interview with camera, microphone and screen share.

You choose the length, add your own questions, and can set a typed task. The report gives an overall score, separate scores for technical depth, communication and integrity, quoted strengths and concerns, and one of four recommendations from “strong hire” to “do not proceed”. Live audio and video are processed to run the interview and are not stored as a recording. Full details are on the pricing page.

What your candidates will experience

It helps to know what you are asking of people. Our guide for candidates, how to pass an AI interview, describes the experience from their side, and it is worth sharing with applicants who have not done one before.