Will voice apply take over the world?

[00:00:12] Welcome to episode 135 of the Inside Job Boards and Recruitment Marketplaces Podcast. I am your co-host, Steven Rothberg, founder of College Recruiter, and currently suffering from a summer cold. So if it sounds like I've been smoking a lot of cigars, it's not. It's a virus. Peter, good to see you. Summer cold in May, you know, what can I tell you? Well, you can't really get a summer cold in January.

[00:00:36] That's true. You can get a summer cold in January, especially if you're living in the Southern House here. Okay, fine. You got me. Okay. Anyway, good afternoon. We usually do these in the morning, but today it's afternoon for us. Who knows when you're listening? Good morning, good afternoon, good evening, depending on where you are. And thank you for joining us. I'm Peter M. Zollman of The AIM Group.

[00:00:59] And what prompted us today to talk about VoiceApply was the acquisition a couple of weeks ago. Tom Chevalier and his startup Tink acquired another startup, relative startup called ApplyCall, which is a system whereby people call and give their information conversationally.

[00:01:27] And the next thing you know, their conversational information has been processed through AI and Tink's AI then gives that information to the employer, etc.

[00:01:39] And we've seen a few of these companies that do the voice apply. What struck me about Tink and ApplyCall was that they are targeting specialized high-end niches or niches, specifically nursing and healthcare.

[00:02:01] And the company that we reported on a couple of years ago, and you and I talked about this, Stephen, but I couldn't find it in our archives. I know it's out there. They were specializing in blue collar. They started in Europe and they said, truck drivers don't have a resume.

[00:02:19] And even if they do, they don't have the time, the patience, the inclination to upload the resume, to apply for a job through a resume. So they call us or we call them and we say, we ask you six questions. We take those six questions and we process the answers to those six questions and we stack rank them and we have a field of candidates.

[00:02:47] So the question for you, Stephen, is everybody going to be doing this in five years? Oh, hell no. End of podcast. It's been a pleasure. Yeah, it's good as always. To be specific, I don't want anybody to think that I feel like this stuff has no merit and that nobody will be doing it. I do think that some will be doing it, but I do not think of this as a replacement for job posting ads or the way that a lot of job boards do business now.

[00:03:15] I see this as being an extension of what we're doing. It's a way for employers to tap into a group of candidates that might not be inclined to go to a traditional job board keyword search, find postings, maybe get matched by AI, resume, upload your resume, it'll match you against job postings.

[00:03:38] What I've thought of with these voice based systems, though, and I think you touched on this, Peter, is that they've largely been confined to what some people call blue collar work, what some people call standing jobs. But it's the customer service, it's the truck driver, it's the day laborer kinds of roles.

[00:04:02] Now we're starting to see this creeping into more white collar sitting kinds of jobs, the jobs that you do at a desk, behind a computer, including a lot of the health care positions. It'll be interesting to see how well it works. Well, obviously, saying to the guy who runs the high volume hiring podcast, there's some real merit to the option of saying we're going to make a thousand phone calls.

[00:04:32] Now, you have to be careful in the States, at least, about do not call lists. And certainly in Europe, there are all sorts of privacy restrictions. But we'll make a thousand phone calls or we'll make a thousand texts or 10,000 or 100,000 to people and say, here's an opportunity you may be interested in. Answer the phone, call us at, and just answer five simple questions or six simple questions.

[00:04:58] And if you're looking for nurses and you can get 200 of them to call you out of 10,000 and you have a story about that, so I'll let you tell the story. The pace of change in the world of business is fast and that drives incredible complexity for the world of human resources. Welcome to the HR Data Labs podcast, a series of conversations with experts inside and outside the world of HR on how to innovate, measure and evolve our practices.

[00:05:28] Our goal is to help provide you with practical examples of how HR has to change to meet the complexity of the business environment. Every week, we'll talk to new and different voices on all aspects of HR. And at times, we'll get irreverent, silly even, and sometimes geek out on the data and technology that underlie the processes that drive the world of HR. But the conversations are always insightful and fun.

[00:05:55] So please, enjoy the HR Data Labs podcast. You were playing softball pitcher and throwing me a nice underhanded pitch there. I was at a conference in mid-May. One of the presentations was a talent acquisition leader at a big healthcare system, along with the CEO of a HR tech or TA tech vendor that offers this kind of voice system.

[00:06:21] The healthcare system was concerned that in nine days, their nurses, 2,000 or 4,000 or something. Yeah, 4,000 nurses were going to go on strike. And if you're running a chain of hospitals, you can't really close the doors like you can if you're running a cupcake shop. So they brought in this vendor who did massive outreach to nurses.

[00:06:45] And the goal was to basically bring in 2,000 nurses to break the strike or to be working while the permanent nurse employees were on strike. They didn't end up needing to do that because the strike was settled. But they played some recordings of the interview process. It was, I would say the best word I can describe it is awkward. The typical AI pauses, the sycophatic language. That's really interesting.

[00:07:13] You know, and just that kind of stuff, I think, made most of the people, most of which were job boards, leaders in the audience, just kind of feel like this technology is coming, but it's not quite here yet. It just didn't come across in a very candidate-friendly, polished manner. But that doesn't mean that it's not going to be there in six months or a year. And it doesn't mean that it doesn't have merit today. I just think it's going to get a lot better.

[00:07:43] We did it. You and I did a quick scan of our archives at the AIM Group. And we found Fiverr is using some of this. And Eleven Labs is funding some of this. Hackajobs is using conversational AI. Don't know if that's audio. I wanted to look it up. And there are at least several other companies doing this service.

[00:08:06] And the one that most impressed me said, you know, dial-less number, truckers, people who are day laborers, people who are, you know, restaurant workers looking for a quick fit. And within 10 minutes, you've been interviewed, six questions. One of the things that struck me about it was that it was language neutral.

[00:08:30] So if you spoke English, if you spoke French, if you spoke Spanish, if you spoke Portuguese, it would translate it and it would stack rank it. It's being used by Tesco or one of the big UK supermarket chains. And they said a ton of their hiring for their entry-level jobs.

[00:08:54] And those entry-level jobs, unlike entry-level jobs in white collar, those entry-level jobs are never going to go away. They're going to be people stocking shelves or straightening shelves, even if there are robots doing some of it. They're going to be people, they're going to be some cashiers at the supermarket, even if most of the cash registers have become self-service. Yeah, and speed matters greatly with those roles, right?

[00:09:23] I mean, if somebody can just sort of be impulsively looking for a job, they're a nurse and they're happily employed, but something better comes along or they're out of work and they're looking for a cashier position, whatever. There are tons of different options there. But if you're Tesco and you can put a candidate through the application screening process, extend a job offer to them in a couple of days,

[00:09:48] you're going to have your choice of candidates when your competitors are taking two, three, four weeks. So it's one of the reasons I like these systems. One of my concerns about them, though, is that along with a lot of the AI-powered hiring system, there's really no way of knowing how well they work. And what I mean by that is if a thousand candidates come in through the process and the AI tools,

[00:10:15] which are almost all built on top of OpenAI, some of them Grok, Anthropic, Gemini, you know, whatever, the big foundational models, all of those companies have said publicly that there is no reliable audit trail. You cannot ask Gemini with any degree of accuracy how it arrived at a decision. It will give you an answer. It'll tell you this is what it did.

[00:10:44] But it's kind of like me saying to you, hey, Peter, remember a week ago when you went for lunch? What did you get for lunch? Why did you choose to buy to get that rather than something else on the menu? Did you like it? To the best of your ability, you're going to give me an answer. You're going to tell me what you think is correct. But you won't actually know it's correct because it's looking back. And I'm concerned about the bias.

[00:11:10] I'm concerned about these systems making the hiring manager believe that this candidate is well-ranked and this other one is poorly ranked. Two things, though. Number one, there's a certain level of verification and authenticity if you have a candidate doing a voice interview. Yes. That's number one. Number two, if I have a big banquet tomorrow night at my hotel and I discover I'm 12 servers short,

[00:11:40] I don't really care about the long term. I'm not really concerned about the bias. I want 12 servers for tomorrow night. And these voice application tools or these voice qualification tools and interview tools can certainly go a long way.

[00:12:02] If I need a trucker who is going to drive for my company and I can qualify him or her through a series of six questions, do you have a commercial driver's license? Have you driven over-the-road trucking or do you prefer local trucking? Do you have a hazardous materials certification?

[00:12:25] And six qualifications knock out 70, 80, 85 percent, leave me with a much tighter pool. And then the AI tool says, here's a list of 24 versus going through a thousand resumes. Most of them never would have come in at all. I think this is definitely an option. I do want to chime in with a couple quick things, though, if you don't mind. No, I don't mind at all.

[00:12:55] Of course. I think what you're describing there is improved efficiency for the employer. And I agree. But that doesn't, improved efficiency does not mean that it's legal. And some of the examples that you stated, I think, would be from the kinds of organizations or the kinds of hiring needs that the employer would be more likely to say, I don't really care what the law is. I just need to be staffed. Where for a hospital system that's been around for 100 years, that'll be around for another 100 years,

[00:13:22] the likelihood of being sued or criminally charged for illegal hiring practices is definitely much more of a concern than somebody who is just throwing a wedding. Speaking of a banquet, I want to know, Peter, why I wasn't invited to the party that you're hosting. You were invited. You just didn't come.

[00:13:44] And last but not least, I want to express my condolences on the disappearance of your hockey team, the Minnesota Wild from the Stanley Club. Oh, yeah. And we're recording this the day after our team was eliminated. And so by this time that people are listening to this, it will be old news. I knew that, but I didn't have, you know, I can express. But sad news. I can express condolences whenever. Thank you, Peter. Thank you, sir. Have a good one.