00:20:36 · Daniel Chait
[00:20:36] S0: Yeah. So, you know, we've described it as the AI doom loop, which is the idea that, you know, in the sort of last few years, you've had everyone on both sides of hiring using tons of AI, but they haven't been given the tools to make things better. So candidates facing a downward job market, a tough job market, have been using AI to apply to more and more jobs. And secondly, to give themselves a shot at each job to customize their content to that job. So if I'm gonna submit a resume to apply for a job as a candidate, I'm now using AI to read the job description, take my resume, and merge my resume to look more like the job description it's looking for, put in all the right keywords, and send it in, which sounds smart, and I don't begrudge individual candidates doing that. But when everybody does that, what ends up happening is everybody's pipeline fills up with more and more applications because they're all getting sent out automatically, and they all start looking more and more of the same. And so I've heard it described as you're trying to find a needle in a stack of other needles. And so for a company, what they end up doing is now they open up a job, and they'll get many hundreds or even thousands of applications for that job, and they all start to look the same. And so they're stuck with this problem of the signal. Like, I don't know where to start. I'm not gonna read 2,000 resumes. I'm not gonna talk to 2,000 people. What do I do? And so they also turn to AI and say, Build me an AI filter. ChatGPT, tell me which 25 people I'm gonna call. ChatGPT will certainly happily tell you 25 people to call, but is that the right 25 people? Is it introducing or reinforcing existing biases? Is it actually looking beyond the resume and finding out anything important about the person? So the answer is no. And so that then deprives the company of needed talent. The other nine seventy five people that are never getting a look, probably the right one's in there, and it's making the job market harder for those applicants, so they respond by sending out more resumes, and the cycle continues. So that's sort of the place we've been in. And if you're in the job market as an applicant or as a hiring manager, you certainly recognize that it's not easy right now. It feels bad.
00:22:43 · Daniel Chait, Guillaume Moubeche
And so what we've done is tackle that problem head on and say, This is not a sustainable solution. AI has so far made things worse, not better. We need to fix this. And so voice AI is a piece of the puzzle, but basically what we said is, if we can find ways for candidates to give better information about what they're good at and what they want, that'll help. That'll help those candidates stand out and get a fair shot at a job. It'll help the companies identify the right talent to focus on. And so we're looking for all kinds of ways to surface better signal, more trustworthy signal to help fix this problem. And sort of my hope is that we can play a meaningful role in sort of unwinding the AI doom loop and helping make jobs and hiring better. [00:23:24] S1: And down the line, have you already like sort of becoming some sort of marketplace of talent? Because if you start like interviewing, you know, like with Voice AI or any AI setup, like thousands or hundreds of thousands or millions of candidates, you eventually like gather a lot of data about specific people. And let's say that they apply for company A, you ask them like a set of questions, etcetera, but you realize like they're not gonna be a fit, however, they might fit for a company like C, D, or E. Are you gonna push that company, that profile or not? Like, is that something you've considered or or not I
00:24:06 · Daniel Chait
[00:24:06] S0: think you're onto something. I would think about it slightly differently. Like, first of all, just marketplace to me is kinda like, I don't Yeah. Yeah. Yeah. Is like kind of where you're selling, like, meat and toasters. Like, I don't think of people. Like, these are humans. And, like, behind every number at Greenhouse, there's a human being trying to get a job, and I think a lot about that. And so... But I think what you're saying that I do agree with is the hiring process has some fundamental assumptions built into it based on the way we've always done things that I think we can reimagine. I think we can sort of question, why do we have resumes and job applications? And these things are like, you know, most... Like, resumes are still on a white rectangle this shape because we used to print them out on paper and put them in envelopes. Like, we don't do that anymore. And so why do we have these limitations, and how can we rethink it? And so if you step back and you're like, oh, this process could work so much better, I think one of the ways that you're talking about that I very much agree with is candidates are putting in a ton of effort into trying to get a job. They're answering a bunch of questions. They're approving their skills. They're choosing which companies they're interested in and which ones they're not. They're doing all this stuff. Can't we make all that more useful to more candidates, to more companies? Can we make the process give better ways to stand out and showcase who you are? Better ways to find among the billions of online profiles, who's the set of people I should be talking to if I'm hiring someone? I think there are some real basic assumptions in how that's always worked that we don't have to be so caught up in anymore.
00:25:31 · Guillaume Moubeche
[00:25:31] S1: And what would you say are like the three big bets that you are making on where the hiring market is going?
00:25:39 · Daniel Chait
[00:25:39] S0: Yeah. Well, I mean, let's take the example that we already started talking about with voice. Like, one of the big problems in hiring has always been resume review, always. The idea that you send in resumes, whether it's in envelopes in the old days or in a website recently, you submit your resume, and then you hope you get an interview. And why do have to hope so hard to get an interview? It's because giving an interview is expensive. A company has to take an employee and put them on the phone with a candidate for a half hour or an hour. Companies are only going to do a few dozen of those for each job. But when hundreds or thousands of people are hiring, that means almost everyone who applies to a job isn't getting a look. And so by questioning that assumption, saying, Why do we have to do a resume review before you get an interview? It's because interviews traditionally are manual and time consuming and expensive. But if we can do them automated in a way that's transparent, that feels good, we can give them to everyone. And so our customers using Voice AI are now saying, Everybody who applies gets a conversation. And so Voice AI will automatically open up the funnel to anyone who applies. And that's great for job seekers who are willing to spend a half an hour answering some questions, improving their skills, and a chance to get ahead. It's great for companies who can now find diamonds in the rough, not care about resume keywords and everything. But from a systematic standpoint, and why do I say we can make things radically better, is because resume review is one of the most biased parts of hiring, always has been. There was a study done in 2003 that showed a 50% racial bias in who gets a resume callback, who gets an interview callback based on the name at the top of their resume. If your name sounds stereotypically black, you have a 50% less chance of getting an interview than the same resume with a name that sounds stereotypically white. And that was twenty five years ago. And so if we're gonna skip a resume review altogether, and anybody who applies can have a conversation, we eliminate completely. We just define that 50% out of existence altogether, and we skip that whole thing. And so that's what I mean by there's real opportunities here to say, we're not just gonna minimize the harm, we're actually gonna make things radically better.
00:27:45 · Guillaume Moubeche
[00:27:45] S1: And do you feel like some people or like some candidates are gonna not truly like the process of being interviewed by an AI? Or like, what's your view on it? Like, do you feel like overall it's taken, like, pretty positively by candidates? Or