Everyone Switched to Async Video Interviews. The Completion Rates Tell a Different Story.

Async video interviewing became the consensus "smart" upgrade to phone screening over the past three years, and the market reflects it — the global async video interview market crossed $1.2 billion in 2025 [1]. The problem is that somewhere between implementation and results, the completion rates showed up, and they're inconvenient.

What Completion Rates Actually Look Like in Practice

Industry data on async video completion rates ranges from 40% to 60% depending on sector and role level — meaning between 40% and 60% of candidates who start an async video interview don't finish it [2]. For high-volume roles, particularly in logistics, retail, and hospitality where candidate drop-off is already a risk, the numbers tend toward the bottom of that range.

This is rarely discussed openly because it creates an uncomfortable arithmetic problem. You put 500 applicants through an async video screen. 250 don't complete it. You now have a shortlist drawn from 250 people — not 500 — and you have no idea whether the 250 who dropped off were unqualified or simply unwilling to record themselves answering prompts at 11pm on a Wednesday.

The standard response is to assume completion rate correlates with motivation. If they didn't finish the video, they weren't serious enough. This logic is almost never examined with actual data. Most teams don't follow up with non-completers to find out who they were. The candidates just disappear, and the assumption covers the gap.

The Experience Signal Problem

Async video introduces a specific failure mode that phone screening doesn't: rehearsal. A candidate who records their answers on a platform that allows retakes — most do — is giving you the version of themselves they've constructed over multiple attempts, sometimes with scripts in front of them. You're not evaluating how they think. You're evaluating how well they prepared.

That matters more as AI preparation tools become increasingly accessible. A candidate can now prompt an AI to generate model answers for any behavioral question, review the script a few times, and deliver a polished 90-second response that accurately represents nothing about how they'd actually perform in the role. The assessment tool is yielding false positives at a higher rate than it used to — and the assessors rarely know it.

Phone screening had its own limitations, but the spontaneity constraint was a genuine signal. What a candidate says in real time, under mild pressure, without preparation windows between prompts, is actually informative. Their communication style, the way they handle an unexpected follow-up question, their level of genuine interest versus rehearsed enthusiasm — those things are legible in conversation and largely invisible in a polished video clip.

How Asendia AI Changes the Equation

Asendia AI is a voice-first AI recruiter that conducts live spoken screening conversations 24 hours a day, 7 days a week. Not recorded prompts, not chatbot flows, not async video. A real-time voice conversation that adapts to what the candidate says, follows up on inconsistencies, probes for specifics, and produces a qualification summary that reflects what the candidate actually communicated under live conditions.

The completion rate dynamic works differently because the participation bar is lower. A candidate who would have abandoned a multi-prompt video sequence at 9pm can answer a phone call in the same time slot, have a 12-minute conversation, and complete a full qualification screen without setting up lighting or re-recording an answer. Voice conversations typically see completion rates above 85% — because picking up the phone and talking is something people already know how to do [3].

What comes into the ATS afterward isn't a video file for a recruiter to watch. It's a structured qualification summary: key candidate statements, role-specific criteria scores, and a ranked shortlist. The recruiter doesn't review 300 video responses. They review 40 screened, documented candidates — and they can do that in an hour. For agencies handling high-volume campaigns, this is the difference between adding a headcount to manage throughput and not having to. Asendia plugs directly into existing ATS workflows, so there's no new system to manage, no parallel dashboard to check.

If you're thinking about what this means for the numbers you're tracking — completion rate, screen-to-interview ratio, pipeline velocity — the post on recruitment KPIs in a post-AI world covers which numbers actually reflect pipeline health and which ones are obscuring it.

Final Word

The shift to async video was a reasonable response to a real problem: too many applicants, not enough recruiter time for phone calls. The completion rate data suggests it partially solved throughput and created a different problem — lower signal quality, higher dropout, and a screening layer that sophisticated candidates can now game in ways that weren't possible two years ago. The question isn't whether async video is better than nothing. It usually is. The question is whether it's better than the alternative, and the alternative has changed. Voice AI that runs 24/7, adapts in real time, and talks to every applicant the same evening they apply is not a futuristic scenario. It's what teams that are no longer losing candidates to async dropout rates are running right now.

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