Why Better UX Won't Fix Async Video Screening

Async video screening was a reasonable answer to a real question. If you have hundreds of applicants and a few recruiters, how do you screen everyone without spending recruiter hours on each one? You send every applicant three to five prompts, let them record answers on their own time, and review the recordings when you can. Every candidate gets screened and nobody's calendar suffers. It's easy to see why the format spread so quickly between 2022 and 2025.

What's harder to see from the employer's side is where the cost went. It didn't disappear. It moved to the candidate.

Think about what the invitation tells an applicant. They've just applied, nobody at the company has looked at their application or spoken to them, and the first thing that comes back is a request to perform for a camera. Most people don't mind the technical part. What bothers them is what it implies: that the process was built around the employer's convenience, and that there's no one on the other end who can respond. The candidate forms a quick opinion of the company at that moment, and it's rarely a flattering one.

People with other options act on that opinion. Someone who is employed, or has three other processes running, will close the tab rather than record a video that may never be watched. The drop-off figures quoted for async video run from 40 to 60 percent before completion, and it's reasonable to assume the people leaving skew toward the ones with alternatives. I've written about that selection effect before, so I won't repeat it here. What interests me more is why the obvious fixes haven't worked.

What the vendors fixed

Async video vendors know about the drop-off, and they've responded the way software companies usually do, by improving the product. Prompts got shorter. Recording works better on phones. Some platforms let candidates do a practice take before the real one. These are sensible changes, and they probably help at the margin.

But they're aimed at the wrong complaint. They make recording easier, and difficulty was never the main objection. The main objection is that the candidate is talking to something that can't talk back. A shorter prompt on a better mobile interface is still a one-way performance. So I'd expect polishing the format to lower the drop-off a little without touching what causes it.

Meanwhile the numbers downstream give employers little reason to look closer. When a recruiter reviews the people who finished the video, the completion data looks fine, the shortlist looks fine, and conversion through later rounds may look healthy. None of that tells you anything about the people who left before anyone evaluated them. The ceiling on the eventual hire is lower than it looks, and nothing in the dashboard says so.

There's also a slower cost building up. Several US states have passed or are considering laws that require employers to disclose and get consent for AI-analyzed video interviews, and the EU AI Act puts added requirements on video tools that draw behavioral inferences about people. Teams whose whole first round runs on video analysis will be carrying that overhead for a long time.

What removes the tradeoff

Async video became the default because there was nothing better for high volume. The honest version of the deal was that you gave up a real first conversation in exchange for screening everyone. That was a deliberate tradeoff, and for a while it was the only one available.

Asendia exists because we think the tradeoff is no longer necessary, and we made it a phone call on purpose. What candidates object to in async video is that nothing talks back. A spoken conversation run by voice AI fixes exactly that: it can hear an answer, ask a follow-up about it, and respond when the candidate asks something in return, none of which a recorded prompt can do. Asendia phones each applicant soon after the application comes in, evenings and weekends included, and runs a structured qualification conversation that adapts to what the person says. It asks about their background, explains the role in concrete terms, and answers their basic questions. The recruiter gets a ranked set of candidates with verbatim excerpts and notes against the role's requirements, written into the ATS. That's what async video promised, everyone screened quickly without a recruiter on each call, except that the candidate has someone to talk to.

I'd expect completion to be much better, for a simple reason. A phone call isn't a new interface. Everyone knows how to take one, and it resembles what a good recruiter would have done anyway. If you want the wider argument for AI that runs pipeline steps itself instead of assisting the people who run them, I made it in a post on agentic recruiting.

I don't think async video will disappear. There are roles where a recorded answer is useful. But as the default first step in high-volume hiring, I think it's on its way out, and not because the companies that built it made bad product decisions. They solved the employer's throughput problem by handing the cost to candidates, and in a market where good candidates have options, that was an expensive place to put it. The first contact is where a candidate decides how committed to be, and that decision follows them all the way to the offer.

Ready to transform your hiring strategy? Schedule a Demo with our founders today!