Why Candidates Finish Live Two-Way AI Interviews
Candidates finish live two-way AI interviews more often than one-way video screens because candidates book their own slot and the interviewer talks back. The evidence is customer-reported: one SDR team interviewed 640 of 640 applicants, and a healthcare staffing firm saw drop-off fall from 58% to 12%. This page lays out those numbers, the mechanics behind them, and the limitations of the data.
What do the deployment numbers show?
Customer-reported completion outcomes from published case studies all point the same direction:
- Healthcare staffing (Canada, 3 provinces): candidate drop-off fell from 58% to 12% after switching to self-scheduled live AI interviews - nurses applying after night shifts could finally interview at 2am. [Customer-reported]
- Denver SaaS SDR hiring: 640 of 640 applicants completed a live AI interview - the entire applicant pool, not a resume-filtered sample. [Customer-reported]
- EdTech (US/Canada/UK/Australia): time to first interview dropped from 4 days to 45 minutes across 4 time zones - candidates interviewed while intent was still hot. [Customer-reported]
- Healthcare staffing: time to first interview fell from 5 days to 3 hours; the 5-day wait had been the point where most candidates disappeared. [Customer-reported]
Why do live two-way interviews complete when one-way videos don't?
Three mechanics explain most of the gap. First, self-scheduling removes the waiting: the recruiter opens a slot window and the candidate picks any slot in it, so nobody waits on a recruiter's calendar while their attention moves on. Second, a live conversation is an interview, not an audition: candidates consistently abandon one-way tools that ask them to record monologues to a silent camera, while a two-way AI interviewer that listens, reacts, and asks follow-ups is treated like a real interview. Third, there is exactly one step: an emailed link, a picked slot, a conversation - no app installs, no multi-day back-and-forth.
Where does the 40-60% one-way baseline come from?
The 40-60% range for one-way video interview completion is the range we cite across this site for asynchronous, record-yourself screening tools; it reflects what hiring teams switching to The Cognitive report about their prior tools rather than a study we ran. We publish it as a directional baseline, and the limitations below apply to it doubly - if your one-way tool completes above that range, the honest comparison is your number against the deployment figures above.
Methodology, sources, and limitations
Figures on this page are customer-reported outcomes from production deployments of The Cognitive, as published in the linked case studies, or arithmetic computed from those published figures and from our published pricing. They are not a controlled study.
Limitations, stated plainly: completion here means invited candidates who finished an interview, but individual customers may define invited differently (all applicants vs. resume-passed applicants). The deployments span different industries, role types, and volumes, so the figures are not directly comparable to each other. There is no control group - customers changed process and tooling at the same time, so scheduling flexibility, interview format, and novelty effects cannot be separated. The one-way baseline is directional, not measured by us. We publish these numbers because they are what our customers actually report, not because they are experimental evidence. Last updated: August 7, 2026.
Frequently Asked Questions
What is a good completion rate for an AI interview?
There is no audited industry benchmark, so compare against your own funnel. The best customer-reported result on this page is 640 of 640 SDR applicants completing a live AI interview; one-way video screens are often cited in the 40-60% range. If your completion rate is low, the usual culprits are multi-day scheduling delays and a record-yourself format candidates resent.
Why do candidates drop out of one-way video interviews?
Candidates drop out of one-way interviews because the format asks for effort without reciprocity: recording monologues to a silent camera feels like an audition, there is no ability to ask questions, and invitations often arrive days after applying. A live interview at a self-chosen time removes all three frictions - which is what the 58%-to-12% drop-off change in our healthcare case study reflects.
Are these completion figures independently verified?
No - they are customer-reported from production deployments, and this page says so explicitly. The case-study numbers (640 of 640 SDR applicants interviewed, drop-off 58% to 12% in healthcare staffing) are each individual customers' own reported outcomes. We publish methodology and limitations precisely so readers can weigh the evidence for what it is.
More research
Deep AI Interviews (why live two-way) · Healthcare Staffing Case Study · Denver SDR Case Study · What Is a Screening Interview?