Ep 54: The Misadventures of AI in Recruiting with Jeremy Lyons - Part 1

[00:00:00] Welcome to Elevate Your AIQ, the podcast focused on the AI-powered yet human-centric future of work. Are you and your organization prepared? If not, let's get there together. The show is open to sponsorships from forward-thinking brands who are fellow advocates for responsible AI literacy and AI skills development to help ensure no individuals or organizations are left behind. I also facilitate expert panels, interviews, and offer advisory services to help shape your responsible AI journey. Go to ElevateYourAIQ.com to find out more.

[00:00:28] Hey everyone, welcome back to Elevate Your AIQ. I'm Bob Pulver and today we're diving into how AI is completely changing the hiring game. My guest is my friend Jeremy Lyons, co-founder of RecOps Collective and a leader in recruiting operations. Yes, RecOps means recruiting operations for those

[00:00:56] not familiar. We're tackling a huge challenge in hiring right now, which is AI-powered mass job applications. Candidates are using AI to apply for hundreds of jobs at once. Recruiters are drowning in applications. We've got fake candidates. We've got all kinds of stuff going on that we'll dig into. But what does this mean for the future of hiring? Jeremy and I break down the good, the bad, and the ethical gray areas in AI recruiting. And we talk about whether AI should be allowed in job interviews,

[00:01:23] why resume screening AI isn't always getting it right. If you're a job seeker, recruiter, or just AI curious, you're going to love this conversation. Let's dive in. Hello, everyone. Welcome back to another episode of Elevate Your AIQ. My name is Bob Pulver. I'm your host. And with me today is Mr. Jeremy Lyons. How are you doing today, Jeremy? Jeremy Lyons I'm doing all right, Bob. How are you doing? Jeremy Lyons Doing all right. Jeremy Lyons Doing all right. It's Friday. There's a beer in my future.

[00:01:49] Why don't you just give my listeners a little background about yourself and what you're working on? Jeremy Lyons Yeah, definitely. So my name is Jeremy Lyons. I am the co-founder of RecOps Collective, which is a company that I founded at the end of 2022 to help make recruiting operations work accessible to companies in all different fields, not just big tech. And around

[00:02:15] the same time too, I started a newsletter specifically for the recruiting operations space, as well as, as soon as OpenAI released the ability to create GPTs, I created the most popular recruiting operations GPT. And last year did the first recruiting operations conference. So definitely keeping myself busy. Also keeping myself very abreast at all the new technologies and how AI can be used,

[00:02:42] not used. How do we create human and AI interactions? So I keep myself pretty, pretty busy, but pretty, pretty plugged in. Even if I look to get out of this space as well too. Jeremy Lyons I think that's great. I think that's, people need to follow your lead. I think that's going to be one of the themes of this conversation, which is, you know, be curious. Always looking to see how you can prove the way you see things, view things, get things done, et cetera.

[00:03:12] Jeremy Lyons Well, before we even jump into that, the thing that I will say is, you know, not necessarily following my lead, but certainly I've made a lot of mistakes so other people don't have to make those mistakes. And I can probably, you know, point things out in different directions for people. Yeah, no, that's, that's right. We got to learn from each other, right? So the event that you ran, remind me, was that a virtual event or in person? Jeremy Lyons First in-person, large-scale recruiting operations specific conference, actually.

[00:03:42] Awesome. So how did it go? Jeremy Lyons You know, people who attended had a fantastic time, a lot of really great information, a lot of really great insights. One of the things that people really enjoyed was it was now attaching faces to names. You know, it's very easy in the LinkedIn-iverse or really anywhere else that's digital to feel like you have relationships with

[00:04:05] people, even if you've never met them in person. And being able to have those conversations in those spaces where you could see how, you know, somebody's mannerisms work and how they're probably delivering what they've been writing. Very, very interesting to see. I think the other thing too that I, I always tell people is comically is like, I'm not a very tall human being. I'm all of

[00:04:31] five, six. But if you were to see me on video, I look a lot taller. So I always tell people if they're going to try and find me in a crowd, don't look up, look down because I'm probably running around in there. But it was also really funny to see some other people who are very tall versus, you know, versus my impression of seeing them. It's funny you say that because when I've met folks at some of these industry conferences,

[00:04:58] I really had no clue, right? So I kind of, I guess I had to look up and down. I will say, just in general, I agree with you. I mean, people, they'd love a true human interaction in real life. You know, I think it enhances relationships. You get to know someone obviously a lot better, whether it's, you know, during a happy hour or just conversing in between sessions and things like

[00:05:23] that. And so I think the power of, of community really sort of comes to life when you get to attend those in-person events where possible. Yeah. I mean, to, to kind of even build on that, I think the, the, one of the greatest, let's call it kind of things that is sad is that people oftentimes don't get to be privy to all of the conversations and some of the most brilliant ideas. I'll call it like the greatest graveyard in

[00:05:51] the world is actually the notes app for pretty much everybody because you're writing down these ideas that are, are never going to see the light of day or need to wait or need to marinate or need to be involved. And while those ideas are great ideas, they sometimes just go buried. You know, you forget about them, but really those in-between conversations are where you learn just so much more than sometimes people being on stage and talking.

[00:06:18] I often think about how AI can be, obviously there's a lot of AI powered sort of note taker solutions. I use them all the time. It just seems like even with that, if it's an interview, then it's definitely going to get, you know, looked over at least the, you know, the summaries and takeaways and key insights. But for a regular call, I mean, it just seems like there's a lot of

[00:06:43] potential, you know, ideas and thoughts that could be collected and mined and aggregated because otherwise, to your point, it just goes off into the ether. I mean, it happens every day, but it just seems like that would be a pretty strong use case for, feed that into, into an incubator and see what spits out the other side. And my brain goes two ways with that. One is, you know, I know Apple intelligence for one has now added

[00:07:09] a feature where you can click a button and it will automatically record the call. At the same time, I take another viewpoint, which is if we, if everything was recorded and then fed back into the machine and stuff like that, are we eventually going to have a Nixon Watergate type recording thing where somebody says, hey, Bob said X, Y, and Z. And it was sort of meant to be a private

[00:07:38] conversation. And now you're sort of feeding AI more information about you. And, you know, if we're building agents and the idea would be to have like sort of the agent that is you, that you train, that you don't necessarily have to keep going back and plugging. It just sort of comes along for the ride with you. Is that information that you will have enough time to go back through, pull out and sort of say like, hey, look, you know, I was exploring an idea.

[00:08:04] It's not necessarily core memory for me and it should not be core memory for you. That's totally fair. I think you've got to, if you were to do it, you've got to put some significant, you know, parameters around what you're listening to, attribution back to individuals. I mean, this, this came up all the time when we were doing social media analytics, even 15 years ago, listening to Twitter or whatever, you're using it in a positive way or a negative

[00:08:31] way. If you can mine it in general or patterns or trends or whatever, it's harder to do sort of anomaly detection. Certainly when it comes to just, you know, brain dumps and off the cuff, you know, thoughts that make it their way into a tweet or something like that. It's not necessarily actionable in itself, but it could be, you know, a breadcrumb of some sort. Definitely. Well, and I mean, I think, I think we are getting really, really close to a point

[00:09:00] for this, but I do think that I see it coming up and I don't understand necessarily how it's taken this long, but I do think that if a company is recording the interview, there is no reason for the company, only the company to hold the transcript of said interview. When I think about the tools that are out there, especially around interviewing, you know, I think of the meta views, the bright hires

[00:09:26] doing that. But one of the AI tools that I love to use for just about everything is Fireflies because it goes back both ways. So you can get the transcript, I get the transcript, and now we are sort of this mutually accountable party. But I can also go back if you were to say decline me and see like, okay, I'm going to do a performance analysis of myself. Maybe I went on for a few longer minutes than I

[00:09:54] needed to for this answer. Or you know what, this is where I can see in the interview for my own performance, because interviewing is performative, that I should have done better. Or they asked me this question and I misinterpreted what the ask of this question was. And if I were to get this question again, this is how I should have answered it. As opposed to leaving the interview and having anxiety about the things that you said, because you can't remember what you said 29 minutes into

[00:10:24] a 30 minute conversation, that now, you know, what has that narrative changed for you? I don't know for sure that nothing goes back to the candidate, but I totally agree with you. There should be some learnings from that experience. I know I've gotten feedback when I've done like, not really an AI interview, but use like a, like a sepia kind of tool, answer these rich,

[00:10:51] three rich text questions, answer them at your own pace. This last time I did it, it was pre-generative AI. So it was certainly, there was no risk of, of me, you know, using AI to answer those questions. But point is I got within minutes, I got response with some constructive interpretation of how my answers were, you know, the inference from my answers. And then, you know, maybe a little bit of constructive feedback and it was greatly appreciated and

[00:11:20] improved the candidate experience. Even if I didn't move further in the process, I got, I got something out of it. Right. And so I agree with you, something should go back. So if they're not doing that, you know, hopefully they will, I wouldn't give them the exact same information where, so you use Fireflies, I use read.ai. So same thing, both are, everyone who attended the call gets that, the summary, the action items, et cetera. And so maybe it's a little bit different,

[00:11:48] just like with sepia, like the dashboard on the backend for the TA team, it looks quite different than what I got as a candidate, but, but still there's, everyone gets some value and insight from the conversation of which we were both a party too. Yeah. Well, and I think that there's, I'm, I'm looking forward to the day where it can actually go like one step further and it might already exist.

[00:12:13] But I think when we, when you look at how communication is and how humans use communication from time and continuum is that so much of human communication is body language and how are you delivering information? How are you proceeding to tell people things? And I think that what would be really interesting from these tools is not only evaluating the quality of the answer, but also

[00:12:41] your mannerisms. And I certainly know as somebody who is neurodivergent, one of the things that's really interesting to me because I've received the feedback is, well, executive presence was, is it distracting to have me move around or be looking around in a conversation as I'm trying to gather my thoughts? Well, that doesn't necessarily mean I'm like looking over at another screen to like try

[00:13:09] and feedback and answer. Maybe, you know, I'm looking up into the left and that has a whole different thing in body language sciences. But for me, the ability to use my body to communicate my message, I think is very important. But to certain organizations, what they kind of want is that stoic, not call it stoicism. It's not necessarily that stoicism, but it's like they want somebody who's going to stand there, you know, shoulders back, chest up, does not move at all because that's the

[00:13:38] stakeholders that this person is going to talk to. And they want to see that as opposed to something else. And that's something could then be fed back and sort of said like, hey, this is the personality of the company. They're not really a company that talks with their hands. So don't use your hands when you talk. That's something that you can learn. That's something that can be taught. And that's also good feedback. I mean, think about how TED talks look and sound. Everybody talks the same, does the same

[00:14:07] thing with their hands. That has become a specific style. And there are people who coach people on how to do TED talks. So I think that we're maybe steps away from that. But I do think that like, as a candidate, that feedback can also be very important to you, especially if your goal is to move up in an organization and be an executive. And to some of us, that's not always the case. Sometimes you just

[00:14:33] want to be really good at what you do and stay at one level. And you don't need to be coached to have this sort of persona that it is not you. No, that's totally fair. I will say I have had a guest on that has a startup that does coaching based on physical presence. She started the company. She was a former professional ballet dancer. And so she got into

[00:14:59] movements and all these things and then somehow got into AI and built this whole thing where, you know, yeah, you can do, you know, full body, you know, recordings and it will coach you about interpretation of your movements and, you know, mannerisms and styles like that. It's pretty, pretty interesting. I mean, being somebody who did ballet and has a lot of dancer friends, I can tell you a lot of dance is

[00:15:25] pain and you have to make it look graceful and not have it show on your face because honestly, being in pointe shoes is not comfortable by any stretch of the imagination. And yet you have to sit there and do that smiling. I would take mannerism coaching from a ballet dancer nine times out of 10. Excellent. She'll be happy to hear that. Rachel Kosser is her name. I want to circle back to

[00:15:50] your, the RecOps collective and some of the feedback that you're hearing about, you know, AI in, in recruiting and talent acquisition. You know, a lot of surveys about folks in this space, pretty wide range of concerns, some optimism, but I guess cautious optimism. Some people embracing it.

[00:16:17] Some people nervous about losing their jobs. Some people think they're irreplaceable because of the human touch that they offer, et cetera. What, what kinds of stuff are you hearing directly from, from your community? Yeah. I mean, the biggest thing that I'm hearing right now is how are we as a community going to overcome basically. Before we move on, I need to let you know about my friend, Mark Pfeffer and his show,

[00:16:47] People Tech. If you're looking for the latest on product development, marketing, funding, big deals happening in talent acquisition, HR, HCM, that's the show you need to listen to. Go to the Work Defined Network, search up People Tech, Mark Pfeffer, you can find them anywhere.

[00:17:44] Application increases. So candidates using AI to apply out to hundreds of jobs. So how do you deal with that volume? The other thing that I'm hearing a lot of is around like fake candidates. So, you know, somebody does kind of the first interview for you and then the next one, you know, somebody else joins or maybe you're paying a service that does all the interviews for you and then you show up day one and

[00:18:11] you're not the person that they've been talking to. So that's certainly a conversation that's happening. And then kind of the third one that I'm hearing a lot of people talk about is a using AI during the interview. And, you know, is that a good thing? Is that a bad thing? You know, it's I think the debate that I hear or at least the common theme that comes out is

[00:18:37] basically something very human, which is we all get mad at the person who's using the thing the best or the most clever because we didn't think of it first. And so it's like you sort of can't get mad at somebody for using the entire game board to to an advantage that shows creativity, that shows how you can go about doing different things

[00:19:02] in a different kind of way. Yet people are sort of saying, well, that's not fair. You're not playing by the rules, which I think is a very interesting conversation that is being had around, well, all right, how how are we going to incorporate these AI tech into our interview process where it's sort of like this is OK to use and this is not OK.

[00:19:29] Like it's OK to use AI to maybe take the initial prompt and use stuff. But like in the interview, what we want you to do is write the prompt in a specific kind of way so that it's not just sort of, hey, what's one plus one? And it gives you the answer. It's sort of like I have a problem. I have one of this. I have one of that. Let's brainstorm some solutions for how we're going to get to that.

[00:19:57] Yeah. So I have some relatively strong opinions about most of that. Well, I want to I want to hear. I mean, this is a conversation. I mean, let listeners hear hear your perspective on the three things I brought up. Well, first of all, using AI to put your best foot forward in terms of the resume and a cover letter if you're actually putting cover letters together.

[00:20:25] Those are tasks that would be outsourced anyway. There are professional resume writers and cover letter writers. That's been a cottage industry for decades. So it really comes down to is this task or is this request something that I'm being evaluated on

[00:20:49] where there's some expectation that this is a work product from my own brain and my own hands, as it were. So I don't have any issue with that. And the fact that someone created use technology to create a mass supply thing that was inevitable. And it's just a reaction to the game that has been created.

[00:21:17] So you're not breaking the rules. You're bending them. But that also means on the other side that someone's going to figure out a way to mitigate that. Maybe it's through two-factor authentication with the applications. Maybe it's some kind of new sort of CAPTCHA kind of thing that says, you know, are you a human being applying to this job? Are you real?

[00:21:45] And do you even know where all these applications are going? So that when you get a response, because the response might be just an acknowledgement that we received it, but the response also might be, thanks for applying. Now go take this, you know, pre-hire assessment or whatever. And now another action needs to be taken place. Now you may build another agent to fill that out too.

[00:22:13] You may have your own candidate agentic, you know, workflow that does all of these things and is able to react when it gets that response and has some of those, some of it's slightly, you know, rules-based, but some of it is just following the patterns and understanding how you would, understanding your voice and how you would respond to those things. Anytime you start, a candidate starts using AI where it should be clear

[00:22:43] and should be common sense where they're trying to understand how you and your human brain works, then that's, I cry foul, right? Like you, it, to me, it reeks of desperation if you can't fill that out yourself. And this isn't about time, the time it takes at this point.

[00:23:09] This is about, let's understand your thought processes because we know you're going to use AI when you get here. Let's hope you will use it to make yourself more efficient and effective. And the company may already have some ways to do that, even if you don't bring your own, you know, AI, so to speak. But, but this is an exercise to understand, especially like for a psychometric, you know, behavioral assessment,

[00:23:34] you know, maybe a take-home assignment that hopefully they're compensating you for your time to do. But, you know, the future is humans plus AI. It's not a zero-sum game, one or the other. So, to me, that's where I draw the line because that is essentially, like you described, sending someone else to take your interview for you.

[00:23:59] I mean, I'm picturing, you know, Ben Affleck, you know, sitting in for Matt Damon and what's it called? Good Will Hunting? Good Will Hunting, my favorite movie that I can't think of the name of. But, yeah, I mean, it's just like, what do you mean? Like, this is not, it's just not right. It's not ethical. There's three parts to that, too, which, I mean, I'm loving that we're having this conversation. The first part, you talked about resume writers as a cottage industry, and it has been.

[00:24:26] I think one of the things that has grown out of this movement in technology has been resume people selling on fear. And that's something that I take offense to, and I know a lot of resume writer folks who also take offense to it as well. And it's this, like, iteration of, I'm going to write you a resume that is ATS compliant or is going to beat the ATS.

[00:24:53] And that is something that, as somebody on the recruiting operations side, as somebody who builds these systems and somebody who, you know, sets up these automations, kind of takes a, like, do you even know? Like, I once, there was somebody who reached out to me once, and they were like, hey, you know, we'll review your resume with our proprietary technology. And do that. And I was like, cool, great.

[00:25:20] And it gave me back some good feedback, and I said, well, what ATS did you use to do this? And the person, like, refused to tell me. And I'm like, that's kind of odd that you're telling me that this is going to beat an ATS, but you can't tell me how you tested these against the ATSs. And I know a lot of ATS companies, and they're just sitting there going, like, you know, like, what?

[00:25:46] Now, a lot of those companies are building in AI, generative AI pieces. Some of them are using keyword search. That's how they're choosing to build. Some are using semantic. Some are using combinations of both. That's just how the technology is right now, but I'm sure that it's going to evolve in things. The second thing is that, you know, people are sort of treating this mass apply thing like it's somehow a new problem.

[00:26:13] And I remember 20 years ago applying to colleges, and this was, like, at the very first iteration of, like, the Common App or the UC App, where you could essentially create one application and fire it off to all of the universities in that system without, with basically just the, like, hey, write a new resume, and it's still going to cost you the same thing.

[00:26:40] And at the time, you know, I didn't, I don't know if this was ever a debate or what people were talking about. All we saw publicly was company or colleges being, like, applications are up 500%. What we didn't necessarily hear at the time was colleges saying, this is how we are looking at resumes now to handle the volume. So this is a problem that has been around for 20 plus years, or if not longer.

[00:27:07] And there hasn't been a lot of research that has come from that area to say, this is how we solved it. Or maybe this is how we didn't solve it, but this is how we mitigated it. And this is how we've, like, gotten through. And then kind of to your last point, you know, the human, the AI, this sort of thing. Brandon Jeffs and Matt Adler had a fantastic conversation earlier this week, week of January 27th that just came out,

[00:27:35] where literally it's, they talked about this idea between AI productivity, human productivity. And when we think about that, you know, you get the John Henry problem, like humans are not going to be able to compete against machines. So what are we going to deliver that is just as equally important from a business outcome? That's not necessarily going to show up on a P&L sheet.

[00:28:00] But how do we sort of talk about this productivity goals and how are we goaling each to sort of reach an outcome? Well, on the productivity piece, certainly I think that's only part of the picture, right? And especially when we talk about, like, individual productivity. I mean, we work in teams. We work in departments and divisions and organizations.

[00:28:28] And so we can't just have a couple random cogs going 10x and everyone else is still at 1x or 2x. And, like, something's going to break. Like, I think that we need to think about, are we effective in our role? Can we use it not just to answer questions or perform some of the routine tasks, but how is it helping us aggregate more information to make more informed decisions?

[00:28:55] How is it finding, you know, signal in the noise? And some of the ways in which it can make us more effective to deliver more value and better outcomes. And so I totally appreciate the focus on productivity because people are almost, they're using AI, but a lot of it is more in the automation ballpark, which, you know, we've been doing,

[00:29:25] or some companies have been doing for at least a decade, you know, things based on, you know, writing scripts and following the preset, you know, rules and parameters. So there's still, you know, quite a bit of that, even if it means putting that into system instructions of an AI agent or something like that. But I do think that, you know, part of understanding, I mean, maybe we should just give candidates assignments where they do have to use AI.

[00:29:53] I mean, it's only a matter of time before AI skills show up in job descriptions. Just like for us, it was, you know, how good are you with, you know, Microsoft Office and, you know, using Mac OS X and, you know, whatever. So that's where I get concerned. You know, you mentioned some of the, you know, you think about students, you know, early career folks entering the workforce.

[00:30:17] This is how you show, like you said, some of that resourcefulness, curiosity and appetite for learning and growing. I mean, these are the durable skills that people are going to need no matter how much AI advances. And so how do we, you know, teach those skills?

[00:30:35] And so while I recognize that some of this could be, you know, resourcefulness, if you're setting up, you know, second screens, trying to, you know, have it give you, you know, answers in the split second, you know, before you can have an uncomfortable pause trying to answer an interview question or, you know, come up with a solution to a particular problem that's been asked of you.

[00:31:03] Or write a particular, you know, script or do some coding exercises or whatever. I mean, it's just, you know, if they, if the vendors and the talent acquisition teams actually have to monitor for potential cheating. I mean, I know this is a life, this is, this is a perennial, you know, cat and mouse kind of game as well, but it just seems like, come on, like, can you, you can be able to do the job when you get here or not. If not, then like have a nice day.

[00:31:32] Right. Well, and I think, so one of the more creative problem solving methods that I've seen to sort of counteract this AI apply, AI tech weaving its way in has sort of been to do something that we've always done, which is, you know, put in a knockout question. And essentially it says, like, it's a text field and it says, if you are an AI, write in the word blank.

[00:31:57] And then all you're doing on the back end of that is saying, if, let's say banana, you know, tag used AI. Or if banana, you know, automatically decline this, what I think becomes slightly harder in that equation becomes, all right. So we know that people are working on AI filters because they still want the humans to apply and, and, and all that.

[00:32:27] So we know that that's the case. But what happens if, for example, you use, you write for your agent, if you see this type of question, bring me back into the loop before you apply. So now you have answers that were written by AI for maybe a portion of it. You're now back into the loop as the individual. You're now writing additional answers. Now you have a 50-50 split.

[00:32:54] So is it sort of going to become something that on a back end filter, you sort of move the toggle up and down that sort of says, if 80% AI boot them or if, you know, whatever it is. But you're right. The ultimate cat and mouse game has always been this employer versus employee fight.

[00:33:20] And what we've seen is sort of when the pendulum shifts between different types of markets, that everybody tries to grab back the pound of flesh that they gave up to get back to like an equilibrium point where they don't sort of feel like it's like they've given up too much. It is essentially the economics game of mutually assured destruction.

[00:33:45] If I have to give up building a nuclear weapon right now, not me talking about nuclear weapons in a violent sense, it's the economics explanation of how this game theory piece came out. You know, it's like prisoner's dilemma. It's exactly what we're running into. I think what is going to be fascinating to me was the job market used to look like it's an employer market. It's an employee market.

[00:34:14] And now we're in this like shift back to the employer market. But in those swings, we didn't necessarily have generative AI at the ability or AGI that will be coming soon at the ability to replace what the people are doing. So now will the pendulum swing back or will the pendulum sort of stop in the middle?

[00:34:41] Because essentially a company goes, well, we don't have to give up these resources to bring people in. We can build smaller teams. Those smaller teams can, you know, have a subject matter expert and agents underneath. And so to your point, AI skills come into things in the same way that like we all somehow needed to know how to use Microsoft Office. That was a skill that still sometimes appears on job description.

[00:35:08] So it's going to be really interesting to see how these things go back and forth. And I can't remember who said it, but it's like there are going to be billion dollar companies that have one person running them because of their ability to sort of understand all of the pieces. And so does that change what our definition of a renaissance person is? In the future, because it used to be, OK, you can play an instrument. You can ride a horse.

[00:35:37] You can talk philosophy and and all of those things, which in a lot of ways, I wonder, you know, we're talking skills based hiring now as being a big thing. Well, is skill based hiring actually the continuation of the you should get a liberal liberal arts degree because you are going to learn all these sorts of things that we're going to teach you how to think?

[00:36:04] And the thinking and the thinking and the creativity and that part comes from like, how can you take everything that you've learned from everywhere and synthesize that into the work that you do? And is that also something which I mean is a whole other conversation unto itself around all of those elements, too? I think there's some. I think there's some.

[00:36:27] Some AIs that are are built now and will continue to be built where you do have that sort of second brain kind of thought where you you can't possibly. But part of it's about recognition. Part of it's about recall. Right. Like, like, you've been working a long time. There's a lot of different things. Some of us are have certain triggers and we can start connecting dots from these different disciplines and different, you know, experiences and things like that.

[00:36:57] Others, you'd have to, like, literally stick it in their face and say, well, what about do you remember when you did this? Do you remember when you did that or whatever? So that's been the debate between inductive reasoning and deductive reasoning people. There are some people where you show them everything and they're like, great. And there are some people where you're like, hey, cool, here are two threads and they're going to go. I've seen this before. It looks like a duck. It sounds like a duck. I'm not going to convince myself that it's a chicken. It's a duck. Yeah.

[00:37:22] No, I mean, I think, you know, those kinds of things will be more useful to certain people than others and in certain roles and contexts than others. But nonetheless, it's incredibly valuable just from the examples that I've seen. And I certainly could use one because I'm not as organized as you. And I just can't keep it all straight or it's just overwhelming.

[00:37:51] The aspects of just going back to the mass supply thing. And I was just wondering if you had, it's more of a technical question, I guess. I'm going to ask it anyway. Anyway, just like you would track through recruiting operations, you would track like source of candidate or whatever you call that. Right. Like whether they came in through LinkedIn.

[00:38:18] Like not just the it could be like the method as well as like the site or whatever. But if you mass supply, isn't there some like unique indicator where some companies would just be like just like some colleges did not participate in the like sort of universal application? Couldn't companies just be like, no, I can already tell there's a, you know, an indicate unique indicator that says this came from this tool. I know that tool is a mass supply tool. I'm not having any part of it. You're going to screw up my whole.

[00:38:48] You know, team and like, no, if you're serious about this job and go to the actual read the job description and and apply. I'm not dealing with this nonsense. Couldn't you identify that and just like block them? I love this question. I'm going to answer it from my from my knowledge base with what I think based on conversations with other people. I will knowingly admit and tell people before. This is a very, very good question.

[00:39:17] And I'm giving my personal opinion based on what I know. And my personal opinion could be wrong. OK, so there's a couple of elements to this question, which is where did the candidate come from? And the other part is how did they apply? Most ATSs right now are not necessarily going to be able to tell you the how they're going to tell you the where. Hey, this is William Tincup.

[00:39:45] And I'd like to talk to you a little bit about Practitioner Corner podcast. It's a wonderful podcast about the journey, the paths of how practitioners, both HR and TA kind of go from high school, college all the way to where they are right now. Some of the things that they've learned, how they've been successful, people that thrive around them, etc. It's a fun podcast. You'll love it. You'll learn from it. Subscribe to it. Thanks.

[00:40:46] And that's because mostly when you look at the where, that is something that we've always, that has always sort of been tracked in recruiting, even when it went back to handshakes and like, hey, you should hire Bob because I know Bob or nepotistic hiring. Like Bob's my brother. You should hire Bob. Here's a handshake. And then somebody says, hey, how'd you find Bob? Well, I'm, I know Bob from Jeremy and Jeremy introduced me to Bob and that's how we got Bob.

[00:41:15] That's been sort of something functionally there. What you bring up though is a really good point, which is you probably could go back into the code base of these things and look at what is in there in the code. And to most companies, well, that could, that might actually start to take more use or like be a better thing.

[00:41:40] Because what you might want to be looking at, and I know that there are things like tracking pixels that you can use that marketing uses. I mean, other tools use these too, where you can say, how long did somebody spend on a specific page? And you could go back and you could say like, it looks like Bob spent five minutes on this job application. If above five minutes, probably human.

[00:42:09] Versus if it's an AI mass supply and it's coming through, you're probably going to see something infantismally fast that you could then say, probably not human. In terms of how this person applied. So I think that it is, but then the question that you have to ask is, all right, we figured this out, but what is the impact?

[00:42:34] Like all data is just data and you can lie with data, which is something I think people sort of think math is this insalibleness of everything. And you need to like it, math is, the math is never wrong. Well, the math can be wrong. It's how you present it. But what's the impact of this? Why do we care? And most of the time, like, well, from the recruiting lens, we care because one, we want to know if this is a real person.

[00:43:03] Two, we want to know how long, what's the candidate experience like? And the application piece, if we're like, hey, look, we're looking at this thing and maybe it's like going through work day and it's like candidates are spending 20 minutes to fill out one application. This is not worthwhile for the candidates. How do we bring this down to a reasonable number? And we're going to call a reasonable number five to 10 minutes. How do we bring it down to that? That's what we care about, the candidate experience.

[00:43:29] But then we have to ask the question is, is ultimately, does the business even care about this? Because how do we present that this person, like most of the time, the business is not going to be like, oh, well, you got this person from shaking hands on the street. What they're going to be going after is, how do we replicate finding great talent for our organization? What does that great talent look like?

[00:43:56] And then address the question of, how do we repeat this? And how is this person driving impact? Ultimately, there are details that matter to our recruiting team that we want to report out because we think that that's great. But the business might not care how you get there or why you got there in that kind of way. So then when I think about all of what I've just said, then it becomes, well, why haven't ATSs designed this as like straight standard whatever?

[00:44:25] Point is, is that they've probably done the business impact assessments. They probably talked to a number of people. And in doing that, essentially determined, you know what? We don't need this. This is a nice to have. It's not a need to have. So we're not going to build this into the platform. It seems like you're getting, I guess it depends on the market, which obviously fluctuates, like you said. Sometimes it's a candidate market. Sometimes it's an employer market.

[00:44:55] But no one has, at least anecdotally, I haven't seen anybody say anything good about mass supply. At least for the candidate, I understand it's not, these applications are time consuming. And we know which ATSs are often creating a lot of that time suck. But, and so I do appreciate it that, you know, candidate time is valuable too.

[00:45:27] And they've been put in this position, right? Because they apply to jobs, they spend the time, they don't hear back, they're annoyed, they're pissed, and they don't know what else they can do. And that includes jobs that you look really well matched for. And you don't, but you don't know. You don't get, like we were talking about the feedback from, you know, the recordings, right, from interviews and stuff like that. Like with the ATS, for whatever reason, and I'm happy to, you know, I guess I'm curious to get your stance on that too.

[00:45:57] But for whatever reason, you apply and they can't seem to, like, get to the point where they can just tell you any reason whatsoever of why you were rejected. So you don't know. You don't know if something got screwed up in how it was received on the other side. You don't know if someone, if a human being looked at it. You don't know if there was something that, you know, ruled you out before it, you know, reached a human being.

[00:46:27] You really are just slept in the dark. And so you're frustrated or whatever. And so, yeah, well, if I could, now I'm just going to play the odds, right? So I'll just fill out this one form and I'll fill it out and whatever. But it just seems, it's certainly not solving a real, the real problem, which is let's skip this whole step. Yeah, everyone that applies, why can't you also give them that response with, you know, hey, we're doing a pre-screen.

[00:46:57] Like, you could automate the recruiter phone screen. Come on. You're never going to talk to that person again, probably. And this is not a part where, going back to what I said before about, you know, I can't be replaced because of the human touch that I offer. I don't, honestly, I've had plenty of recruiter phone screens, even had a couple somewhat recently.

[00:47:19] And I guarantee if I had proceeded, which I did not, like there's very little chance that I would ever talk to that person ever again. So it would sound like we're building a relationship, right? And I wouldn't, if you didn't tell me, that could have been done by an AI and like, I would probably be able to tell, but a lot of people wouldn't. So you're touching on a couple of really key points here. And I'll kind of go through like three of them and I'll circle back and I'll describe each one.

[00:47:48] First one is, is this is where recruiting breaks some sales. The second one is the question that inevitably every recruiter gets asked and comes back to our RackOps team and says, what's the data here? Or how do I pull the data so that I can address this question? And that is, how many number, how, how many people am I going to have to talk to before somebody gets hired? And then the third part of that, which you just talked about, which is the AI interviewers out there.

[00:48:18] I know a couple of companies that are building that come to mind. Brain Trust, use Brain Trust, Talent Llama, number of them are out there. I'll go back, kind of came this way. I'm going to go back the exact same way. To your point, what you could do is give everybody a first round interview if you are using these tools.

[00:48:41] Because essentially then all you are doing is they are going to conduct the interview the exact same way. You're going to basically turn it into the scientific method. You have a hypothesis that is this role. Or the hypothesis is that you're going to find somebody who has applied who has done this role. You're now testing to get back to that.

[00:49:03] So if you can interview a thousand people because you're using an AI interviewer, which can be done at any point in the day. So now you've got when it's convenient to the candidate. It's not going to get tired. It's going to get you sort of the same results. Why wouldn't you then do that? So that's question one. And I think that that's a broader question that is getting answered.

[00:49:29] And usually the counter argument that you hear for that is, well, they need the human touch. Ultimately, they want the job. So now we take it back to the question from the hiring manager, which is, how many people am I going to have to talk to before I find out? Well, the truth is, one, if you know exactly what you are looking for and you know exactly what they're going to need to do

[00:49:57] and you are able to articulate that very, very clearly to the people or the thing that is searching for it, then you only need to have one conversation for that person to be hired. Or, well, not one conversation, but you know that you will only need to talk to one person. The problem in that is inherently a human problem, which is the I'll know it when I'll see it or I'll know it when I get the vibe from it. Okay.

[00:50:22] Bringing it back up, sales and recruiting oftentimes get talked about in similar manners because there's a lot of trailing things that sales does that then recruiting adopts. Sales gets revenue operation sales ops. Recruiting comes back, gets revenue sales ops. It all starts out of this thing where technology is getting more sophisticated. The need to use data to make sure that you're maximizing the amount of time that people are spending is useful.

[00:50:51] And when things are going back and you want to make sure that you're maximizing and not pulling away from resources. Okay. Problem between sales and recruiting is that when you look at the top of the funnel, you want more people in sales. You don't necessarily want more people in recruiting because now you're dealing with the most combustible element on the planet, which is human emotion.

[00:51:17] So you, and you only have usually one role to sell. You don't have more product. It's not, it's not like your recruiters are selling product where it's like, oh, sorry, we ran out of this shipment. We're going to reach back out to you once we've made more and we can ship it to you and your order will be fulfilled. So that is where some of these breaks happen. And one person who talks amazingly about this is actually like Jim Miller,

[00:51:48] who is, talks about how like you don't want to, the purpose of writing a very detailed job description, which is essentially an advertisement and a piece of job marketing and employer brand. Obviously I recommend pay attention to like Katrina Kibben, James Ellis when it comes to employer brand, brilliant people talking about things, very human too. So everybody can understand them are talking about,

[00:52:14] but Jim talks about how you start to change things. When you start to change language and you write a really detailed job description, that will help you with your funnel. And then there are ways that you can also help yourself with your, with your funnel because you don't want to create waste and you certainly don't want to piss people off. And you certainly want to give everybody feedback, but you can't. That's a time constraint.

[00:52:42] Do I think that future tools will be able to sort of say, hey, look, you know, we moved X number of people ahead in this process. All these X number of people that we moved ahead were above this sort of level or had this many years of experience. And based off of your resume, we did not see that. That tells you how you can go back and you can edit your resume to be better. So that's the feedback.

[00:53:08] But what we've also seen is companies avoid giving feedback because of legal cases that have been brought against them for that. So you, it's all of these different pieces and components. And I know I just like laid out like three very sort of different interconnected things. But like, that's part of the human component piece that like, we also all have to figure out for ourselves.

[00:53:33] Because if that's going to go into any sort of level of data that is then going to be used by AI to learn and to grow and adapt and help augment and make us better, you can't start with bad data going in because you're going to get bad data coming out. And that's essentially what will happen. I think that all makes sense.

[00:53:56] I guess, I guess, as without getting into sort of legal hot water, which of course. But that's the best hot water to get into. Come on. Well, that's generally, I mean, I'm not an employment lawyer, but obviously the time I spent in the AI governance space and understanding some of the legalities,

[00:54:18] at least as it relates to using AI or any algorithm in the process, you know, I've learned quite a bit. But it just seems like, you know, for example, if there's adverse impact, you could have adverse impact all day long. But if nobody actually explicitly calls the EEOC here in the US and files a complaint, then it just, you know, it perpetuates, right? Yeah. Jeremy, thank you so much. This has been fantastic.

[00:54:47] I did not expect this to go for so long. And we will talk again soon. Thank you again so much, Bob. This has been a lot of fun as a conversation and as far as things go. And, you know, I'm very excited to see how this comes out. Absolutely. Great. Thanks, everyone, for listening. Really appreciate it. See you next time.