What is an AI interview?

One-way (asynchronous) video interviews. The candidate gets a link, sees a question on screen, and records an answer within a time limit. Nobody is on the other end. A recruiter may watch the recordings later, software may score them, or both. This is the oldest and most common kind.

AI phone screens. A voice AI calls the candidate — or the candidate calls a number — and holds a short spoken screening conversation: availability, pay expectations, experience, a couple of open questions. The transcript and a summary go to the hiring team. These have grown quickly as voice AI has become good enough to hold a natural conversation.

Chat-based screening. A chatbot on the careers page or by text message asks the knockout questions, answers questions about the role, and books qualified applicants into interview slots. Common in high-volume hourly hiring.

Live AI interviewers. A newer category: an AI avatar or voice that runs a full conversational interview in real time, asking follow-up questions based on what the candidate said rather than reading a fixed list.

Most current tools work from the transcript — what the candidate said — and compare it against the criteria the employer set: did they mention the required experience, did they answer the question asked, did they give a concrete example. Some also summarize and rank candidates for a human to review.

Earlier generations of video tools also analyzed facial expressions and tone of voice. That drew heavy criticism over accuracy and bias, and several vendors publicly dropped facial analysis. If you're evaluating a tool, ask directly what it scores and whether a person makes the final decision.

What an AI interview is genuinely good at is consistency: every candidate gets the same questions in the same order, and the first-round logistics — can you work weekends, does the pay work — get answered without anyone's afternoon disappearing into phone calls.

AI interviews are not unregulated, and the rules apply to the employer, not just the vendor.

Illinois — the Artificial Intelligence Video Interview Act requires employers who use AI to analyze video interviews to tell applicants beforehand, explain how the AI works and what it evaluates, and get their consent. It limits who the videos can be shared with and requires deleting them on request.

New York City — Local Law 144 covers automated employment decision tools. Employers using one to screen candidates must have it bias-audited within the year before use, publish a summary of the audit, and notify candidates in advance.

Everywhere in the US — existing anti-discrimination law still applies. If a tool screens out people based on a protected characteristic, or makes it harder for a candidate with a disability to be assessed fairly, the employer is responsible, whoever built the software.

More states are legislating in this area. General information, not legal advice — if you're adopting a tool, check the current rules where you hire.

Usually not. AI interviews exist to compress volume. If you're hiring one front-desk person and twenty-five people apply, the volume problem is an afternoon of phone calls, and a fifteen-minute call you make yourself tells you more than any tool will — particularly for a receptionist, where how someone sounds on the phone is the job.

Where a tool starts to earn its cost is when you're always hiring — several locations, seasonal staff, high turnover in hourly roles — and the screening calls are eating a manager's week.

If you're doing it yourself, the structure is the part worth copying from the tools: the same questions in the same order for everyone, logistics first, and a score written down straight after the call. Our phone screening questions tool builds that script, and the receptionist interview questions page covers the in-person round with a scorecard.

Treat it as a real interview, because a person usually does review it. A few things help:

• Answer the question asked, then stop. Transcript-based scoring rewards relevance; rambling hurts.
• Use concrete examples. "I handled about forty calls a day, mostly rescheduling" carries more than "I'm great with customers."
• Use the words from the job posting where they're true for you. If it asks for scheduling experience, say "scheduling."
• For video: test your camera and microphone first, look at the camera rather than the screen, and use the practice question if there is one.
• For a phone screen: take it somewhere quiet, and speak at a normal pace — voice AI handles natural speech better than slow, over-enunciated answers.
• Ask whether a human reviews the result and how to request an accommodation if you need one. You're entitled to ask.