What is an AI interview?

An AI interview is a first-round screening interview where software asks the questions, by voice or text, and AI then helps the hiring team review the answers.

How an AI interview works

The name covers a few different products, but most follow the same path. A hiring team sets the questions. A candidate opens a link and answers them. Software turns the answers into something a person can review.

A typical run has five steps:

The hiring team picks the questions and the criteria for the role.
The candidate gets a link and is told AI is involved.
The AI asks the questions out loud or on screen, and the candidate answers.
AI reviews the answers against the criteria.
People on the hiring team read the results and decide who moves forward.

The main types

The differences between types matter more than the label. Ask which of these you are looking at before you compare anything else.

Scripted voice: the AI speaks a fixed set of questions, and every candidate hears the same ones.
Conversational voice: the AI writes its own follow-up questions as it goes, so each candidate gets a different interview.
One-way video: the candidate records answers to questions on screen, and nothing talks back.
Text chat: the candidate types answers to a bot.

Scripted or conversational

This is the choice that shapes everything else. A conversational interviewer feels more natural and can dig into a vague answer. The cost is that candidates no longer get the same interview, which makes their answers harder to compare and harder to defend.

A scripted interviewer is less impressive to listen to. It can't chase an interesting detail. What you get instead is a fair comparison, because everyone answers the same questions in the same words. Research on interviews favours that kind of consistency, as our guide to structured interviews explains.

What AI interviews are good for

They help most when the first round is the bottleneck. That usually means one of these situations:

You have more applicants than you can phone.
Your candidates work shifts or live in other time zones, so booking a call takes days.
Different people run your screens, and their notes can't be compared.
You want a written record of what each candidate said.

Where they fall short

An AI interview is a filter, not a hiring process. It can't sell the role, answer a candidate's real questions, or notice that someone has an unusual background worth a second look.

It can also get things wrong. Speech-to-text can mishear people. A score can miss the point of a good answer. That's why a person should read the transcript, not only the number.

Some candidates won't like it. In a Pew Research Center survey of 11,004 US adults run in December 2022, 66% said they would not want to apply to an employer that used AI to help make hiring decisions. Only 32% said they would. Attitudes may have moved since, but the safe assumption is that some good people will need reassurance.

Who should make the decision

People should. The same Pew survey found Americans opposed AI making a final hiring decision by 71% to 7%. Legal rules point the same way. In the EU, the GDPR gives people the right not to be subject to a decision based solely on automated processing where it has legal or similarly significant effects, subject to exceptions.

A useful test: could a person on your team explain why someone moved forward or didn't, using what the candidate actually said? If not, the tool is deciding, whatever the sales page says.

The law around AI interviews

Rules differ by place and they keep changing. New York City's Local Law 144 covers automated tools that substantially assist or replace discretionary decisions in hiring. Where it applies, it requires a bias audit and notice to candidates.

The EU AI Act lists AI systems used to evaluate job candidates as high-risk in Annex III. Regulation (EU) 2026/1744 set the date those high-risk rules apply to Annex III systems at 2 December 2027. Our guide to telling candidates about AI covers the notice side in more detail. None of this is legal advice, so get your own before you start.

What to check before you use one

Ask these questions of any tool, and try the interview yourself first:

Does every candidate get the same questions?
Are candidates told it's AI before they start?
Can you see the evidence for a score, in the candidate's own words?
What happens when a candidate says too little, or the connection fails?
Is a recording kept, and who can see it?
Can someone who can't or won't speak still take part?
Has the tool had a bias audit, if you need one where you hire?

How Vettika approaches it

Vettika is a scripted voice interview. It doesn't improvise and it doesn't ask follow-up questions, so every candidate gets the same questions. Candidates are told it's AI and must consent before they start, and they can switch to typing at any point.

After the interview, AI scores the transcript against your rubric and backs each score with quotes checked against what the candidate said. People on your team make every decision. We keep the transcript, not a recording, and there is no video.

Sources

Common questions

Is an AI interview the same as a chatbot?

Not usually. A chatbot takes typed answers. An AI interview is more often a voice call in the browser, though some let candidates type instead.

Do AI interviews replace recruiters?

No. They can take on the first screen, but people still need to read the results, talk to candidates and make the decisions.

Do candidates know they are talking to AI?

They should. Tell them before the interview starts. In some places the law requires it, and it is the fair thing to do everywhere.

See question templates