The Candidates Who Close the Video Tab

There's a question I'd ask about any screening step before adopting it: what happens to the people who don't finish? For most steps the answer barely matters, because almost everyone finishes. For async video interviews it matters a lot, because a large share of candidates don't.

I understand why teams moved to async video. The problem it solved was real. There were too many applicants and not enough recruiter hours to phone them all, and letting candidates record answers on their own time meant nobody had to sit on the phone all day. On throughput, it helped.

But the completion numbers are awkward. The industry figures usually quoted put async video completion somewhere between 40% and 60%, depending on sector and role level. I'd expect the low end in high-volume hiring like logistics, retail, and hospitality, where candidates were already easy to lose.

Take a simple case. You send 500 applicants an async video screen and 250 of them never complete it. Your shortlist now comes from 250 people. And you have no idea whether the other 250 were unqualified or just didn't want to record themselves answering prompts late on a weeknight.

The usual assumption is that finishing is a test of motivation: if they couldn't be bothered to complete the video, they weren't serious. I suspect it's rarely checked. Most teams don't follow up with the people who dropped out, so they never learn who those people were. The candidates disappear and the assumption fills the space they left. It could easily be wrong. Someone who drops out might be someone with no quiet room to record in, or someone who applied from a break at work and never got back to it, and neither of those tells you anything about how they'd do the job.

The people who do finish have a different problem. Most async platforms allow retakes, so what you see is the version of the candidate they put together over several attempts, sometimes with notes in front of them. You end up judging how well someone prepared rather than how they think.

That gets worse as AI tools get easier to use. A candidate can ask one to write model answers to any behavioral question, read the script over a few times, and deliver a polished 90-second answer that says very little about how they'd perform in the role. So the screen probably lets through more false positives than it used to, and the people reviewing the videos have no easy way to tell.

Phone screens had plenty of problems, but they had one thing going for them. A candidate answering in real time, with no pause between questions to prepare, shows you something true. You hear how they communicate, how they handle a follow-up they didn't expect, and whether their interest sounds real or rehearsed. Those things come through in a conversation and mostly vanish in a polished clip.

So async video solved the throughput problem by giving up two things the phone had: most people finish a phone call, and nobody can retake an answer in one.

This is why we built Asendia as a voice AI that talks rather than records. It phones candidates and holds a live screening conversation with them, evenings and weekends included. A spoken exchange is much harder to script than a recorded clip. There's no retake, and when something doesn't add up or an answer is vague, it follows up and asks for specifics, so a memorized model answer runs out quickly. Taking part is also easy, which matters just as much. Someone who would abandon a multi-prompt video after a long day can pick up the phone, talk for 12 minutes, and be done, with no lighting to set up and nothing to re-record.

What reaches the ATS afterward is a structured summary rather than a video file: what the candidate said, scores against the criteria for the role, and a ranked shortlist. Instead of watching hundreds of clips, a recruiter might review a few dozen people who have actually been screened. That changes what some familiar numbers mean, like completion rate and screen-to-interview ratio, which is the subject of a post on recruitment KPIs in a post-AI world.

Async video is usually better than nothing. The better question is whether it beats the alternative, and the alternative is different now than it was when most teams switched. If you run async screens today, one thing worth doing is pulling a sample of last month's non-completers and reading their applications. If a good number of them look like people you'd have wanted to talk to, the completion rate was measuring something other than motivation.

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