Greenhouse sets sights on AI interviewing as next TA game changer

Hiring platform Greenhouse has agreed to acquire Ezra AI Labs, a startup focused on conversational, voice-based AI interviewing. The deal is set to close later this quarter and signals a shift in how talent acquisition (TA) can address the surge of AI-assisted job applications and the cacophony making it harder to discern candidate signals and fill reqs.

Applications per recruiter on its platform have jumped 412% since 2023, according to Greenhouse. At the same time, fewer than 7% of applicants receive an interview. Recruiters are facing a daunting challenge: more candidates than ever, but less clarity on who’s actually qualified.

“Fundamentally, people want to make the process work. Candidates want jobs. Companies want to make hires, and AI can be a great tool to make all that happen. AI is not the problem in and of itself, but the tools that people have been given haven’t been up to the task,” Greenhouse’s cofounder and CEO, Daniel Chait, said.

Candidates can use AI to optimize résumés and applications. Companies can then deploy AI to filter them. The resulting dynamic is noisier, less trustworthy system where strong candidates can be missed entirely.

The first wave of AI tools hasn’t helped, Chait said. AI tooling in ATSs and hiring platforms “saw the job seekers as the problem.” Solutions created more “friction” for applicants, making it harder.

For Ezra founder Ophir Samson, voice AI can be the unlock, both to reduce the noise, but also to enrich candidate signal.

“The beautiful thing about voice is that when people speak, they share much more about themselves than when they write,” Samson told HR Brew. “If we were going to have this conversation over text message, we’d have 5% of the information communicated. You can get to understand people—whether those people are candidates or the recruiter and the company—in far greater fidelity than you can then when you’re using just a piece of paper like the résumé.”

Instead of screening out candidates based on their résumé, the new feature aims to give applicants a structured, voice-based conversation about the role and their fit for it right at the top of the funnel. Every candidate is asked the same role-specific questions, evaluated against a set rubric, and offered transparency into the assessment process.

“If you think about what an AI interview can do, all of a sudden it really changes the dynamic, because as soon as you apply to a job, you’re now eligible for an interview,” Chait said. “Which all of a sudden means: don’t apply to the job unless you’re willing to spend some time talking to them and taking that first interview.”

Chait said the AI solution “brings it back to the job seeker” rather than simply creating a layer of friction for them. Offering an AI interview creates an entirely new signal.

“Job seekers are motivated to get a job, and so they will do things to get jobs, and if you give them things to do that are productive, that help them showcase who they are to the company, hone in on the right role,” he said, “They’re likely to want to do those things, and when you do that...that increased amount of signal is the way that we break this doom loop.”

Rather than relying on résumés (and thanks to that AI doom-loop, many now are often AI-generated or heavily optimized by AI)—recruiters can assess AI interview transcripts, structured scores, and look at conversational data to understand how a candidate thinks and communicates. It’s even designed to detect overly scripted or AI-generated responses, according to Greenhouse.

For candidates, Ezra aims to help address the invisibility and obscurity often associated with hiring. It flips the hiring dynamic from one of winnowing down the wrong people to a search for the candidate with the right signalling that will be right for the company or position.

Ezra’s voice AI works with hiring managers, recruiters and other stakeholders to generate a rubric for the interview. In natural conversation, Ezra’s voice AI is designed to needle away at generalizations in order to create a precise rubric and flags human inputs that may be illegal or noncompliant so they’re not included—and when the voice is talking too much, you can interrupt.

“A very structured interview in a very structured evaluation process is critical to us, and that is deep into the engineering of the voice system so candidates will always be asked the same questions, whether they’re interviewing at 9am on a Monday or 4pm on a Friday, which are not the case for human interview,” Samson said. “They can interview Ezra back. We’re seeing lots of candidates do that. They ask Ezra a ton of questions about the role or the compensation and the benefits and things which are really important to them to kind of get out of the way earlier. But they feel awkward asking a human recruiter, lest they come across too mercantile.”