Greenhouse Voice AI Quotes Your Answers. Rehearse Each Bullet Aloud

Most of your applications get cut before anyone talks to you. In a survey of 2,832 job seekers, 81% in the U.S. had been rejected without a conversation, and Greenhouse's answer, Voice AI, replaces that resume skim with a spoken first interview that gets transcribed, scored against a rubric, and quoted back to the recruiter.

That changes what you prepare. A strong resume bullet you can't back up out loud is now a risk at the very first gate. I covered the written side in Greenhouse Voice AI Scores Your Transcript. Prep the Resume It Checks. This piece is about the other half: saying each bullet out loud until you can defend it.

What does Greenhouse Voice AI actually do to your application?

Greenhouse Voice AI gives every applicant a first interview with identical questions and criteria, then scores the transcript against the employer's rubric and hands the recruiter the full text plus an explainable evaluation. Your resume no longer decides alone whether you get a conversation. What you say in that conversation decides whether you get a human.

A few facts worth knowing before you take one:

• It isn't brand new. Greenhouse announced it on Oct. 15 as part of its "Orchestrated Hiring Ecosystem," but it brought Voice AI Interviewing to customers in May 2026. One analyst described the October launch as existing products "assembled under a larger pitch."
• It came from an acquisition. Greenhouse bought Ezra AI Labs, closing May 27, 2026. Ezra founder Ophir Samson is now Greenhouse's Head of Voice AI.
• It reaches past Greenhouse customers. Voice AI works even for employers that don't use Greenhouse's applicant tracking system (ATS), the software that stores and routes applications.
• It's U.S.-only for now. Ezra is available in the United States, with other regions planned.
• Named customers include Anthropic, HubSpot, and the NFL.

Greenhouse CEO Daniel Chait framed the pitch bluntly: "Hiring still runs on artifacts from the paper era." Chief people officer Sharawn Tipton put it from your side: "The application used to be a front door. Now, too often, it's a dead end."

Why is an AI screen better odds than a resume skim?

On the available numbers, being screened by AI beats being skipped by roughly 4x. Greenhouse says fewer than 7% of applicants ever get an interview, while 28% of candidates who finished an AI interview moved on to the next round in its May 2026 survey.

That gap is directional, not exact, because the two numbers come from different surveys. But the direction matters. Most candidates fear the AI interview as a new place to get rejected. The data says the real killer was the stage before it.

Being rejected without an interview is the default outcome of a resume-first funnel. Greenhouse also reports that 86% of U.S. recruiters say at least half of applications are AI-written. That is the pile your resume is sitting in.

Candidates aren't sold yet. In the same 2,832-person survey, only 27% said they'd choose an AI-led interview, while 36% would prefer a written application only. Greenhouse's own candidate report found 38% of U.S. candidates have already withdrawn from a process because it included an AI interview, and another 12% say they would if required.

The math argues against walking away. If you withdraw, you're back in the pile where most people never get a conversation at all.

What happens after you finish one

The May 2026 Greenhouse data also shows the downside of the AI interview itself:

Half of candidates hearing nothing is a real problem, and it's the employer's problem to fix. But 28% moving forward is a far better rate than the under-7% interview rate for applicants overall.

What does the Voice AI rubric score, and what does it ignore?

Greenhouse scores the transcript, not your voice. Its product page says scoring is "always rubric-based and based on audio transcripts," and Greenhouse says stutters, pauses, personality, and confidence aren't what gets graded. The words are.

The mechanism runs in three steps:

1. Speech-to-text turns your spoken answer into a transcript.
2. A large language model compares that text against the employer's rubric and ideal-candidate profile.
3. Each question gets a score, the overall score is a weighted average, and a written rationale cites specific things you said.

That last step changes how you should prepare. A vague answer doesn't just score low. It ends up quoted in the recruiter's notes, next to someone else's answer that named a tool, a number, and an outcome.

So most AI voice interview tips about pacing or "energy" miss the point here. What matters is whether every claim on your resume comes with concrete detail when someone asks about it: the numbers, the tools, your exact role, and what happened.

Why does reading an AI-written script backfire?

Reading a generated script aloud invites a flag the recruiter sees first. Greenhouse's Voice AI lists "flags for inconsistencies and AI-generated answers" as a feature, and each flag marks the exact moment in the transcript.

Reporting on the product suggests those flags don't change the numeric score. They don't have to. A recruiter opening your evaluation sees the flag before reading anything else. Add the 86% of recruiters who already say most applications are AI-written, and a flagged answer confirms the suspicion they came in with.

Inconsistency flags carry a second risk. Greenhouse doesn't spell out what counts as an inconsistency, but if your resume says you "led" a migration and your spoken answer says you "helped with" it, that gap sits in a transcript the recruiter reads. Inflated bullets used to cost you later, when a hiring manager pushed back. Now they can cost you in round one, in writing.

The fix isn't to stop using tools. It's to make sure the resume says what you'd actually say. That's the kind of resume Refolk writes: it builds from your own work history rather than a template, so the bullet and the spoken version of the bullet start from the same facts.

Can you tell which employers use Greenhouse Voice AI?

Mostly, no. Recruiter profiles almost never reveal which ATS or screening tool a company runs, so plan as if any application might lead to a voice screen.

In Refolk's index of professional profiles, 97,293 people in the U.S. hold Recruiter or Talent Acquisition titles. Only 120 of them list Greenhouse as a skill, at employers including Meta, Palo Alto Networks, YETI, Paraform, Lumos, and Zearn.

A caveat: these counts show which tools recruiters put on their profiles, not how many companies use Greenhouse. That gap is the point. You can't look up who'll screen you. And because Voice AI works for employers outside Greenhouse's ATS, even knowing a company's ATS wouldn't settle it.

My earlier piece cited Greenhouse data that 70% of U.S. candidates weren't clearly told beforehand that AI would evaluate them. Put the two together and the conclusion is simple: prepare every application as if a voice screen is coming.

How do you prepare for an AI phone screen?

Turn every relevant resume bullet into a short spoken answer packed with specifics, then rehearse it out loud in your own words. That's the whole method. The rest is doing it for each posting.

1. Pull the bullets the posting cares about. Read the job description and mark the handful of bullets on your resume that map to its requirements. Those are what the rubric is most likely built around.
2. Write the facts, not the script. For each bullet, jot down four things: the situation, your specific role, the tools you used, and the measurable result. Use notes, not full sentences.
3. Say it out loud, recorded. Talk through each bullet from the notes without reading.
4. Read your own transcript. Run the recording through any transcription tool and read it the way the rubric would. Count the specifics. If an answer contains no number, no tool, and no named outcome, it'll score like filler.
5. Fix the mismatches. Wherever your spoken verb is weaker than your written one ("I supported" vs. "I led"), change the resume, not the story.
6. Repeat per posting. Different rubrics weight different skills. The bullet you lead with for a platform role isn't the one you lead with for a product role.

Step 6 is where most people give up, because re-tailoring a resume for each application is tedious. That's the work Refolk takes off you: paste the posting, and you get your own resume back rewritten for it, plus a fit score that tells you which requirements you can genuinely speak to and which ones you'd be bluffing on. The bluffing ones are the ones to rehearse hardest, or drop.

What to do in the interview itself

• Name the tool, the number, and your role in the first two sentences. The rationale quotes specific responses, so give it something worth quoting.
• Don't pad. Long answers with low specifics dilute a transcript.
• Ask your questions if the format allows it. In the 2,832-person survey, 33% said the chance to ask questions would make them more willing to take an AI interview.

What should candidates ask employers about the AI screen?

Ask whether a human reviews responses and how you'll be evaluated. Those were the top two things candidates said would make them more willing to do an AI interview: human review of responses (49%) and an explanation of how they'd be evaluated (40%).

Greenhouse reports that 90% of candidates rate its Voice AI experience 4 or 5 stars and that opt-out rates are just 10%. It also publishes monthly independent bias audits by Warden AI on its AI Assurance Dashboard. Those are vendor numbers, so weigh them accordingly, but a published audit gives you something concrete to check.

Questions worth sending the recruiter:

• Will a person read my transcript, or only the score?
• What criteria is the first interview scored on?
• If I'm not advancing, will I hear back? (Given that 51% never do, this one is fair.)

What changes if you're applying outside the U.S.?

Outside the U.S., you get the same resume skim without the AI interview meant to replace it. Ezra launched U.S.-only, and Refolk's index shows 60 Greenhouse-tagged recruiters in the U.S. for every one in the UK.

Yet the rejection-without-conversation rate in the UK and Ireland is 81%, identical to the U.S. Australia sits at 85% and Germany at 86%. Those candidates are getting cut at the skim stage at the same rate or higher, without the voice screen that's supposed to give everyone a turn.

For non-U.S. applicants, the written resume still carries the full weight of that first cut. The rehearsal method still pays off, though, because the first human screen will probe the same bullets.

FAQ

Does Greenhouse Voice AI judge how I sound, like pauses or confidence?

According to Greenhouse, no. Scoring is rubric-based and runs on the audio transcript, and Greenhouse says it doesn't score stutters, pauses, personality, or confidence. A large language model compares your transcribed answers to the employer's rubric, scores each question, and writes a rationale quoting your responses. Spend your prep time on specifics, not delivery.

Can I use ChatGPT to write my answers for a Greenhouse AI interview?

You can use it to organize your notes, but don't read a generated script aloud. Greenhouse Voice AI flags AI-generated answers and inconsistencies at the exact moment they occur in the transcript. The flag reportedly doesn't change your score, but the recruiter sees it first. Rehearse from bullet-point notes in your own words.

Should I withdraw if an application requires an AI interview?

Usually not. 38% of U.S. candidates have withdrawn from a process over an AI interview, but the numbers favor staying in: fewer than 7% of applicants ever get any interview, while 28% of people who completed an AI interview advanced. Withdrawing puts you back in the resume skim, where 81% were rejected without talking to anyone.

How do I know if a company uses Greenhouse Voice AI?

Usually you won't know in advance. Only 0.12% of U.S. recruiters in Refolk's index list Greenhouse as a skill, and Voice AI works even for employers that don't use Greenhouse's ATS. The safe assumption is that any application could lead to a voice screen, so tailor and rehearse your resume for every posting you send.