Video interview software is any tool that uses recorded or live video as part of hiring. The label covers four very different products.

• Live video interviewing replaces the meeting room, usually at final stage.
• Asynchronous video asks candidates to record answers to fixed questions on their own time.
• Video questions with written or spoken answers put the employer on camera instead of the candidate.
• AI video analysis scores candidates from their facial expression and tone of voice.

The important distinction here is which side of the process the camera points at, because that decides who carries the effort and whose appearance ends up in the assessment.

Because they move work and risk onto the candidate while offering them nothing in return. Asynchronous video interviews, also called AVIs or one-way interviews, are the format behind most of the criticism of video interview software, and the objections are consistent.

Candidates need a quiet room, a decent connection, presentable surroundings, and the confidence to talk to a lens with no one on the other end. Recording under a timer adds performance anxiety that has nothing to do with job ability, and it disadvantages people who are nervous on camera, hard of hearing, less fluent, or simply applying from a shared flat.

The employer side tends to not be much better. Review time does not scale, since a recruiter still watches every submission: 250 candidates at ten minutes each is more than forty hours of viewing. And even with no algorithm involved, appearance, accent, and background enter human judgment, which quietly reintroduces the surface signals that structured interviewing exists to remove. It is worth understanding how that bias shows up before adopting the format.

Because two of their needs are legitimate and text alone does not meet them. The first is confirming that the person who did the interview is the person applying, which matters more now that candidates routinely use AI to help them apply. The second is human: hearing someone's voice and seeing how they carry themselves before a customer or audience can tell a recruiter something real before they commit an hour of a hiring manager's time.

Dismissing those needs is why so much advice on this topic goes unread. The needs are fine but it is the standard implementation that is wrong.

Well, not where the scoring infers emotions from a candidate's face or voice. Article 5(1)(f) of the EU AI Act prohibits AI systems that infer emotions from biometric data in workplace contexts, and the European Commission's guidelines state explicitly that hiring processes fall within that context. Exceptions are narrow, covering medical and safety purposes only.

Two details here are important to point out. The prohibited practices have been in force since February 2025, so this is not a future obligation. And while the high-risk obligations covering recruitment tools were deferred to 2 December 2027 under the Digital Omnibus agreement, that deferral does not touch the prohibitions. Any tool that grades candidates on facial expression is off the table now.

It follows seven rules. Together they are the difference between a video question and a video interview:,Only qualified candidates are asked. Video comes after a candidate has already met the criteria, never as a hurdle at the front of the funnel.

1. No preparation required. No lighting, no framing, no backdrop, no rehearsal. If a candidate needs to get ready, the design has failed.
2. It is quick. One or two short questions, answered in the moment, not a recorded assessment.
3. No pressure. No single take, no penalty for pausing or making a second attempt.
4. It is voluntary. The candidate chooses whether to answer on video.
5. No image-based analysis. Nothing about the candidate's face, expression, or appearance is analyzed or scored.
6. No one is automatically disqualified.The video cannot move a candidate backwards. It exists to add context for a human, not to automatically filter.

Read together, the rules describe something modest: short, smooth, and candidate-friendly, yet comprehensive enough for what recruiters actually need.

Yes, and better than an AVI does. A recruiter reviewing a shortlist of already-qualified candidates can see and hear each person for a minute before deciding who to invite, which answers both needs directly. Because no score depends on it, there is nothing to game and nothing to appeal, and because it is optional and effortless, candidates rarely opt out.

Assessment stays where it belongs, on the substance of what candidates said, in text or by voice, scored by deterministic models where the same answers always produce the same score, traceable to the response that earned it.

Hubert works this way by design: interviews can present questions as text, audio, or video, video answers are optional and never scored on appearance, and no candidate is automatically filtered out by one. It is the approach our Candidate Pledge commits us to. If you want to see it in one of your own roles, reach out to us for a chat.