Updated 3 November 2026
A sales candidate can deliver a polished pitch on camera and still struggle when a prospect pushes back. A one-way recording lets someone prepare and repeat an answer, but it cannot show how they respond to an unexpected objection or change their approach during a conversation. Candidate research suggests that some job seekers leave hiring processes that use AI interviews, although no published study gives a reliable dropout rate for sales roles specifically. The format matters because your team needs to assess the conversation, not just the pitch.
When you review completion rates, check the denominator. Completed interviews divided by invitations sent is the invited-to-completed rate. Completed interviews divided only by starts is the started-to-completed rate. The second looks higher because it excludes everyone who never opened the link.
If you invited 200 SDR candidates and received only a small number of recordings, you need to know where the candidates left. Was the invitation unclear? Was the process too long? Did they start but not finish? Or did the format itself give them little reason to take part? Measure each stage before deciding what to change.
The sales-only problem
A pitch into a camera, or an actual objection
Illustrative example, not a real transcript.
A recorded response can be rehearsed and repeated. A live sales simulation can ask a follow-up based on the candidate's last answer. That does not make every live format automatically better, but it lets you assess a skill that matters in sales: responding to the prospect rather than delivering a memorised pitch.
What candidate research says about AI interview drop-off
Greenhouse's 2026 Candidate AI Interview Report surveyed job seekers in five countries. The figures in this section come from its US group (n=1,200, May 2026). In that group, 38% said they had walked away from a hiring process because it included an AI interview, and another 12% said they would.
Asked what pushes them out, they named three things most often.
What US job seekers named as a reason to walk away
Greenhouse 2026 Candidate AI Interview Report, US respondents, n=1,200, May 2026
Greenhouse does not state who the percentages are out of, so read each as "33% named it", not "33% of those who left".
The same survey found that 70% of US respondents were not clearly told upfront that AI would evaluate them, and 21% said they found out only once the interview had started. Your team can address this directly: explain in the invitation that AI is involved, what it assesses and who reviews the result.
It is also not AI itself that candidates are pushing against. Only 19% of US respondents want less AI in hiring. The top asks are upfront disclosure, at 44%, and a clear explanation of what the AI is measuring, at 39% (Greenhouse, US respondents, n=1,200, May 2026).
These figures are US-specific, not a sales-hiring benchmark
Greenhouse's own method note reads: "Unless otherwise noted, all findings cited in this release reflect U.S. respondents (n=1,200)." No figure above is marked otherwise. The 2,950 total includes candidates from the UK, Germany, Australia and Ireland, and those are not in the 33% or the 38%.
The UK cut of the same survey, reported by HR Grapevine, shows the markets differ. Greenhouse's release does not publish the UK sample size, so read these as a comparison, not a benchmark.
Same survey, two markets, two sets of numbers
Source: Greenhouse 2026 Candidate AI Interview Report, May 2026. US is the n=1,200 group. The UK sample size is not published.
The US and UK results differ. Do not combine them into one global percentage or present either as a sales-specific benchmark. When you see a completion claim, check which candidates were counted and how the rate was calculated.
There is no reliable public dropout rate for sales candidates
The Greenhouse report does not break interview drop-off down by job function. A percentage described as sales-specific without a published method and sample should not be treated as a benchmark. Measure your own SDR and account executive funnels instead.
One Humanly case study for Cellular Sales, a nationwide Verizon authorised retailer, reports completion rising from 16% on one-way video to 35% after a format change. The account does not publish its sample size, dates, measurement period or method. Treat it as a company-reported example, not a benchmark or a result you should expect.
Until reliable data is published by role, use the general candidate research as a warning to measure your own process. The stronger case for changing the format in sales is that a one-way recording cannot show how someone handles a live objection.
Why a one-way recording misses a core sales skill
A cold call changes when the prospect says no. The candidate has to listen, understand the objection and respond without losing the conversation. A recording that ends after a prepared answer removes the part of the interaction your sales manager most needs to assess.
Here are the four things a sales req needs to see, and whether each format can show them.
What a sales req needs to see
A recorded format can show how a candidate structures a pitch. It cannot show how they respond when the prospect rejects the first approach, asks a difficult question or changes the subject. That is a different skill, and a live simulation lets your team observe it.
Estimate the impact using your own SDR completion rate
Use the calculator below to test what a change in your invited-to-completed rate would mean. The starting values mirror the Humanly case study above; they are not a benchmark or a forecast.
How many conversations does the format cost you?
Starting values are the Humanly case study's 16% and 35% (a company-reported case study with no published sample or method), not a benchmark. Replace them with your numbers.
Plain arithmetic: invited multiplied by the completion rate. It is not a forecast for any format.
With the example values, the same 200 invitations produce 32 completed interviews at 16% and 70 at 35%. That is a scenario, not a forecast. It does not model the additional recruiter time needed to review more completed interviews. Start with your own invited-to-completed rate, then ask candidates who did not finish what stopped them.
What a live sales simulation lets your team assess
Fabric runs the sales simulation inside its live, two-way AI interview. The candidate makes the call, while the AI plays the prospect and responds with objections as the conversation develops. Your sales manager can review the result and decide whether to progress the candidate. Confirm the current score-editing and audit controls with Fabric rather than assuming scores cannot be changed.
The interview runs in a phone browser without an app, supports weak 4G and allows a candidate to rejoin after a dropped connection. Candidates can reschedule. AI disclosure and recording consent are built in across more than 35 languages. The invitation should still explain the process in plain language. See the candidate-experience guidance for details.
FAQ
How many sales candidates drop out of AI interviews?
No published study reports interview drop-off by job function, so there is no sales-specific figure. The general figure is that 38% of US job seekers have walked away from a hiring process because it included an AI interview (Greenhouse, US respondents, n=1,200, May 2026).
Is it the AI that candidates object to, or the format?
The evidence points to the format and the lack of notice. Only 19% of US job seekers want less AI in hiring, and the top asks are upfront disclosure at 44% and a clear explanation of what the AI measures at 39% (Greenhouse, US respondents, n=1,200, May 2026).
What should we tell sales candidates before an AI interview?
Say in the invitation, not at the start of the interview, that AI will conduct and evaluate it, and explain what it is assessing. In the 2026 Greenhouse survey, 70% of US job seekers were not clearly told upfront. Another 27% named lack of disclosure as one of the biggest triggers for walking away, although the report does not specify the denominator for that breakdown (US respondents, n=1,200, May 2026).
What is a good completion rate for a sales interview?
No published benchmark exists for sales roles, so compare your own invited-to-completed rate before and after a format change. Ask which count any quoted rate sits on: completions over everyone invited, or over everyone who opened the link, which always looks better.
Test a different format on one open SDR role
The quickest way to assess the format is to see the cold-call simulation in action. Fabric's live AI interviewer can run a live role-play in which the AI plays the prospect and responds to the candidate's approach.