The Future of AI in Recruitment with Co-founder of PeopleGPT, David Paffenholz

In this conversation, David Paffenholz, founder of PeopleGPT, shares insights into his journey from education to entrepreneurship, the impact of social media on business growth, and the challenges and opportunities in the recruiting landscape. He discusses the features and functionality of PeopleGPT, the importance of addressing bias in AI, and the evolving role of recruiters in the age of technology. David also highlights the company's growth, market position, and future developments, emphasizing the significance of building a strong company culture in a virtual environment.

BTW for the keen listeners, here's a link to the David's go to album by Peter Fox.

Creators & Guests

Host

Martin Hauck

Co-Founder of The People People Group & Host of From A People Perspective

What is From A People Perspective?

A podcast about fascinating professionals, how they got to where they are and where they’re going from the lens HR, Recruitment and People Operations hosted by Martin Hauck.

Marty (01:50.459)
All right, people, people, it's time to talk to David Paffenholtz, founder of PeopleGPT slash Juicebox. David has been at the cutting edge of AI driving, recruiting forward with Juicebox slash PeopleGPT.

And it's a platform that really I've used personally. I saw it go viral. Your post on it go viral. That's originally actually when I wanted to interview you, I think I reached out a few times and we just never connected and that was like 99 % my fault as well. But I'm glad PeopleGPT keeps coming up in the People People group, which is the community that I run.

And I've used the tool myself. It's awesome. So I'm really grateful to get a chance to chat with you.

David Paffenholz (02:50.904)
Thanks so much for having me on. I'm excited for this.

Marty (02:53.733)
Yeah, no. So yeah, you just had a quick high level. Yeah, so you've been through Y Combinator. Previously, you were at Snapchat and I was looking through your LinkedIn profile as a half decent podcaster should do and saw that you graduated from Harvard in three years.

David Paffenholz (03:15.584)
That's right, yep.

Marty (03:16.625)
Cool, cool. Yeah, I guess like give us a quick walkthrough of like your career history, what you've been up to, and then we can dive into the day to day.

David Paffenholz (03:27.704)
For sure. So yeah, I went to college. I'm actually originally from Germany. So the US was kind of, or me coming to the US was through college. And then, you know, COVID happened in the middle of that. And so that was really the time where I started getting into the tech world, startups, and also starting to learn about the talent and recruiting space. And so at the time I interned with a venture capital firm, NEA, and I worked with their talent team.

Marty (03:48.986)
and

David Paffenholz (03:55.266)
And it was actually a really interesting way of getting to know the talent space because one, I got to see kind of the internal recruiting that was happening, but two, also what the portfolio companies were working on and the different challenges that the different TA leaders were having through that role. And so that was my very first exposure and also really where my interest in the space first came from. And then kind of in the years thereafter, as we started thinking about starting the company,

Ishan, my co -founder and I ultimately said, hey, you know, this is the right time for us. We want to go out and build this. I had just graduated college and then we were fortunate that YC accepted us as well to be able to kickstart the business and really go out there. And so, yeah, that was the original journey of how we got here.

Marty (04:40.953)
Awesome. in terms of just curious, did, you in YC before the post one? Cause from my knowledge of you came through that post. So were you in YC at the time or was it pre YC?

David Paffenholz (04:54.508)
Yep, so we did YC one year before the post. And so it took us almost a year to develop the first version of the product. Then we launched it. And then that's now roughly one year ago. So we did YC two years ago, first post one year ago.

Marty (04:57.649)
Okay.

Marty (05:05.979)
Gotcha.

Marty (05:09.937)
Okay, so I'm saying the post. So for the folks that are listening, like, I feel like it was like a pretty big thing because I feel like, I don't know, how many likes did you end up getting or how many like?

David Paffenholz (05:21.102)
It was around like 8 or 9 ,000 likes, which translates to a lot of LinkedIn impressions.

Marty (05:23.567)
Yeah. Yeah. Yeah. Hundreds of thousands. Yeah, that's awesome. And I imagine that was a pretty transformative moment or was it just like, yeah, it's another day.

David Paffenholz (05:35.822)
certainly. We had no clue how that launch day was going to go. we were thinking, Hey, you know, we're going to be happy if 200 people try out the product. And then suddenly we have this crazy LinkedIn launch and, we're up all night trying to make sure the platform's still running and like everything starts breaking. And, it was, it was not designed for, for the skill that we had on, on that weekend alone, we had 30 ,000 people tried out. And so it was really exciting. It was really exciting to see, you know, people are keen on seeing how AI can impact the space, like what solutions can be out there.

Marty (05:38.609)
You

David Paffenholz (06:05.768)
but also kind of showed us a clear direction of like what we still needed to do, what we still needed to develop, what problems we were not solving yet and where the product needed to

Marty (06:17.509)
Yeah, that's fair. So just interesting kind of side question here, just given that circumstance, but in terms of your take on, I guess, social media and content and whatnot, if you go through that experience, you probably have a pretty biased opinion towards it. But how has that really impacted your business? And also in terms of maybe framing a little bit more broadly, as opposed to being specific to like

folks that are building products, but even just people at companies or recruiters at companies trying to get their brand out there. Like how impactful was like a viral post.

David Paffenholz (06:57.462)
Yeah, for us, it was definitely what kind of kickstarted our company. I'd say overall, especially within the HR space and the recruiting space, the platforms like LinkedIn or social content just make such a huge impact in one, building your credibility and two, being able to kind of get out there through word of mouth and be that for like personal branding, for company branding, for really any direction. It's just such a high impact, low cost.

Kind of like an obvious thing to do in my opinion, even if it requires getting over that initial kind of that initial factor that initial hesitation of like, hey, I'm not sure about posting something or like, I don't know if anyone wants to read my LinkedIn post, etc. Putting it out there and then taking that first step in my view is super important.

Marty (07:41.563)
Any

Marty (07:45.227)
just get into the habit basically.

David Paffenholz (07:47.936)
Yes, yeah, I had to force myself initially to like now we do I do like two or three posts a week from my personal LinkedIn and then we do one or two from the company one and it's been a good habit and they keep on getting more engagement. It's like different directions fun way to experiment as well.

Marty (08:02.885)
Nice, nice. One of the things I do on the podcast is I usually do icebreakers and we skipped over that entirely, but I feel like it would be wrong to do to not do that. So first things first, and this is a of a custom one, but is there any chance of you calling it people people GPT?

David Paffenholz (08:21.134)
Maybe I like the ring of it.

Marty (08:24.881)
All right, cool. A little bit more seriously though, in terms of your music taste, we always kind of like, as you can tell, I'm a bit of a music fan. If you had one album to listen to or one band or a musician, what's your go -to? The only person you'd like, you get one and that's it. You can't listen to anything else.

David Paffenholz (08:52.622)
okay. This is going to be niche because it's German. As mentioned, I'm originally from Germany, but there's a German artist called Peter Fox in German. He's my go -to. Like since for 15 years, there's one album in particular that I keep on listening to over and over. So it's definitely niche. But if you speak German or are interested in German music, it a shot.

Marty (08:57.521)
Yeah.

Marty (09:15.323)
That's nice, nice. I'll have to get the link from you to share in the link so maybe we do an intro. No, that's probably a bad idea, but no. And all right, midnight snacks. If and when you do midnight snacking, what are you gravitating towards?

David Paffenholz (09:19.31)
You

David Paffenholz (09:30.858)
ice cream. There's this place called Salt and Straw here in San Francisco that I only moved here like six months ago and it's just the best ice cream ever. So if I have that in the freezer, that's gonna be gone.

Marty (09:44.805)
You're in a good place. Yeah, no, my ice creams. My wife, my wife buys chocolate mint, mint chocolate ice cream specifically because like that's the only flavor I won't touch. But anything else is like there's, there's some right now that I know about and it's just like sits in my brain. Like I know it's there and I'm doing my best to, yeah, yeah, no fellow ice cream enthusiasts. This is going to be a good pod. what was your very first paid gig that you can remember?

David Paffenholz (09:53.219)
Hmm.

David Paffenholz (10:13.046)
first paid gig that I can remember was tutoring in ninth grade.

Marty (10:21.199)
Okay, nice. What subject or is it just broad?

David Paffenholz (10:25.582)
Back then, I think it was math. So I did some math and then economics and kind of the later grades.

Marty (10:33.403)
Nice. Interesting. yeah, so there's, it's interesting. Like we talked, we started off obviously based on, on where I know people GPT from and what you're working on from, but there is this whole other aspect of what you're working on and it has nothing to do obviously with the social media side of things and everything to do with like you making recruiters lives easier, you and your team and everyone there making recruiters lives easier. And I having used the tool, it's insane. Like,

And oftentimes I'm like, you know, there's a founder. talked to a bunch of founders on a regular basis of HR tech by nature of the community and whatnot. And I'm like, like the one thing sometimes that I kind of warn against is when folks haven't really been like integrated with like that function. Like it's like, I've never done recruiting, but I'm building a recruiting tool. so I didn't necessarily see it there. I'm like, my God, like awesome. Then you have a viral, but if you don't truly understand the problem, you can't really solve it. So.

Maybe you can kind of like share a bit more and touched on it, right? You got to see firsthand like the recruiting side, but maybe just like, I'd love to hear your take on like some of the problems that you saw and why you came up with the solutions that you did.

David Paffenholz (11:45.336)
For sure. So I think overall fully agree with what you mentioned. It's really hard to build in a space if the founders don't have like one deep understanding, but then in my view, even more importantly, like a deep empathy for solving the problem that's being solved. Like you really have to think from the user's perspective, why is this something that actually helps or actually matters? And so I mentioned it took us like a year from when we started YC to launching the first version of that product. Of that year, the majority of that time and the first six months entirely,

was just spending time with recruiters and TA leaders and shadowing them. And so we're basically just trying to learn as much as we possibly can, what are the daily pain points? Where does a lot of time go? And where do potential frustrations arise in the day to day? And so we're very grateful. There was a lot of people who were very generous with their time in allowing us to, one, sharing what they were doing and two, allowing us to kind of see and really get a firsthand insight.

And so we kind of started off in with, with different companies that we, that were people that we had in our network. And so it was a lot of kind of venture backed companies, but then we very consciously tried to go beyond that as well. And so really making sure that one, were consciously choosing a problem and not just kind of stepping into something. And then two, making sure we have a good understanding of what, how the user actually feels about things. And so that was how we got started. And the kind of reasoning that we had from there is that.

the, especially in a recruiter's job, a disproportionate amount of time still goes into sourcing. And the actual methodology of that sourcing has not changed much in the past 10 years. so LinkedIn recruiter has become the default option. LinkedIn recruiter is great. It has a ton of data, but it essentially still acts as a database. And so the whole kind of LinkedIn recruiter or Boolean search method is about creating filters to quickly search through that database, which by nature is one, a non -ideal way of finding people.

Marty (13:17.104)
Yeah.

Marty (13:39.352)
Hahaha

David Paffenholz (13:40.29)
because you'll end up with so many people that somehow match the filters, but don't actually match what you're looking for. And two, it's very user time intensive. So you're going to be keeping on doing that for hours and hours. And you still can't be confident that there's not someone you haven't missed out on. So not only is it time intensive, but it's also imperfect and prone to errors. And so because that seemed like such a frequent and repetitive use of time, we said, hey, let's try to build something better.

And we think right now is actually the time to try it because we have this new technology unlock through large language models.

Marty (14:14.577)
Awesome. And for the folks that don't know PeopleGPT, maybe just like your high level, what is it and what does it do? Because I know we just jumped into it. We're doing things a little bit squirrely here today. That's all on me. But yeah, can you give me a quick, can you give everyone a high level overview of PeopleGPT?

David Paffenholz (14:34.028)
Yep, People GPT is an AI powered people search engine. So we help recruiters find the time that they're looking for. And we also have components of the platform that let you get data about that talent and let you engage that talent through emails.

Marty (14:48.793)
Yeah, yeah, having having used it, I think that's kind of half the problem is like I've used it myself. So like, there's less, there's less need for an intro. But for in terms of

more on the high level side of things like having worked with companies like snap and you're you're a founder yourself now like has there been like a big shift in your mindset between like working for an organization and you know being sort of like the one in charge and the one building out the product yourself.

David Paffenholz (15:26.286)
good question. So I think the, I think one noticeable difference for us was that the kind of in the previous experiences we've had, we always were kind of, or I was a member of a larger team and like that team itself was a small team within a large organization and snap, snap, isn't that huge of a company, but it's a, it's a pretty big size company. think we were around 5 ,000 people when I was there. that number could be a bit off.

And so it felt like there was a little bit of a disconnect between the kind of individual actions that we were taking on a day to day and how that was like influencing the future of the company. And then very much the flip side when, when starting off and building something and kind of being a founder of like, Hey, you know, if I don't do something right now, nothing's going to change in the company, nothing's going to happen. And so that was a big mindset, mind mindset shift, and also quite humbling. Like there was so much that we had to do zero to one and so many things that went wrong, that.

Marty (16:10.747)
Yeah.

David Paffenholz (16:21.742)
can really help to change the way I think about things.

Marty (16:25.521)
It's not it's always tricky to sort of any habit or any like mindset and you like to change that is always tricky. So it's good that you realized it. I was listening to one of your earlier podcasts and you talked a bit about founding it virtually and like that's that's a that's a unique experience that not many folks kind of get get to go through has how is that process?

influence the culture for you from your perspective.

David Paffenholz (17:00.75)
That's a good question. So we found it initially virtually largely due to COVID and kind of the constraints that that imposed on us. And then when we did YC, that was also partially virtual. And so YC started doing their first events in person again that that batch. It was in the summer of 2022. habits that we got into by being virtual initially, I think were a lot of transparent and frequent communication.

Marty (17:07.076)
Mm

Marty (17:10.606)
You

David Paffenholz (17:29.652)
especially between my co -founder, Sean and myself, we try to keep each other in the loop on what we're doing, even if it doesn't directly impact what the other person is doing, like on an hourly basis, if not even more frequent. and so that can sound like it's excessive, but it helps us because we're super aligned. We don't want to spend time on something that ends up being not important. and I think we've carried that through now to our in -person culture. So we're mostly in person here in SF now. but we were, we try to be very open, and.

Marty (17:37.36)
Hmm.

David Paffenholz (17:59.48)
kind of encourage frequent check -ins even in ways that would be like maybe more frequent than usual.

Marty (18:05.285)
Yeah, yeah. In terms of skills or habits that you've developed earlier on in your career, are there any that have still sort of stuck with you today that you you just don't let go of because they've they've served you wells or so well over the years?

David Paffenholz (18:23.148)
I try to ask a ton of questions, especially in areas that I don't understand or I'm not familiar with. And that I think helped me early on because it kind of exposed me to a lot of new things. And I think that helps me a ton today because I get to learn more and more. then so the recruiting space itself is so massive. Like there's so many different niches and types of recruiting that I get to learn every day with every, every customer I speak to something new. so the, it from in terms of industries or even the types of searches.

Marty (18:29.199)
Mm

David Paffenholz (18:51.379)
from what it's like to run a really effective executive assistant search to thinking about what market mapping in advance looks like before even a search kicks off. And so there's so many day -to -day things that I didn't know before that I get to know every day.

Marty (19:06.395)
fair. One of the things that is on any recruiters mind or any HR person's mind when it comes to AI, obviously, because it's a it's a touchy subject to a certain extent. And HR is the one that's kind of going into like, privacy matters and protecting individual information and all that fun stuff. And from a recruitment perspective, there's always the challenge of like bias, right? I guess

And you've talked a bit about it in the past, but how do you see AI sort of like evolving to ensure that there's fairness across a wider array of like decisions that should be made by humans, but we're offloading decisions to AI to help us so we can get to the more complex things or the things that need more nuance, right? So curious to get your take on that because you're literally in the AI trenches.

every day when it comes to sourcing and recruiting. I would love just like a broader take and sort of that high like more focused sort of thoughts on like bias and how it's evolving.

David Paffenholz (20:15.682)
Yup. I think for any AI system, the most important part that the creators of it can do is have a really close eye on the data that goes into it. And in the end, the data that's used to train it is what will create any bias or cause any bias afterwards. so making sure that the data set that's included is one, a very high quality, but two, that you're really aware of what is actually in that data. And especially like nowadays where the volumes of data are just so massive.

Marty (20:41.167)
Hmm.

David Paffenholz (20:44.374)
it can become really hard to control that. And so I think that's the most essential piece. I think in our use case specifically on the sourcing side, there's a lot of control that we're fortunately able to have. And so in terms of profile data, for example, we actually remove the majority of a profile's information before we use it to train any models. And so that includes removing names, removing any unique identifiers and more. So that the actual information that goes into it is purely experiences of an individual.

Marty (20:46.0)
Yeah.

David Paffenholz (21:13.422)
with as much information removed as possible. Now there's still some imperfections in that process, but it's a lot closer than to what it should be.

Marty (21:22.885)
That's awesome to hear. I didn't know that. I I imagine that's somewhere on the like features page or in the, not everyone reads the whole like read me of an application. You use it, they jump in, they start using it, it works great. But like, you don't even know some of the stuff that's baked in. So that's really cool to hear that that's been baked in. I've got sort of like a side tangent on that that I'll share with you. And it makes me feel like a boomer because this was like pre -AI, but about.

I was about 10 years ago now, myself and a few developers at a company I used to work at, we were just kind of like talking about how shitty bias is in the first place. And it's just like, it's unconscious. you know, there's at first it's unconscious and then it becomes conscious and you're like actively just promoting these unfair systems. But so he, the, the one developer was like, I could probably fix this like over the weekend. And so he developed like a Chrome extension called unbiased. If I.

And so at the time I was using lever as the recruitment tool and, and I was using LinkedIn, obviously primarily as the place that I was looking for people's profiles. So you could toggle and just like turn off the, the name, but like to know that you've already just baked that in, right. It was kind of funny because one of the, like, one of the, one of the ATS is that we implemented the tool for it. Cause it's just like an overlay, right? So you just.

eliminate if you can find the person's name regularly on any website. So we do, we did it for Twitter and, LinkedIn and whatnot. And so, one of the ATS is that we had was just like, we're closing deals because of your Chrome extension and it's free and thank you so much. And that it was like an open source project at the time. so that that's been always like really near and dear. So it's cool to, cool to see that you've, you've baked it in one of the like hardest things.

to hear as a recruiter from candidates, especially in Canada and foreign or North America, when you hear about a person's like they're trying to get their first job in the country. And then they're like, well, this is my name, but do think I should like anglicize it so that it's more, you know, they understand that the bias exists there. And it was just like, I was always hearing that story from, from a lot of the, the, the new immigrants that I was talking to who are fantastic candidates.

Marty (23:45.627)
but didn't necessarily have that Canadian or North American experience. And so that's kind of where it was born out of. And so I guess curious, like what drove you to include that aspect of, you know, trying to eliminate bias in, in, to, your product.

David Paffenholz (24:04.406)
Yep. Yeah. So the, I think the simplest way for that, that we think about it is we want to create the best kind of recommendation or the best results for your search and, any kind of objective ranking of, of, being able to match experiences as well as possible should by default not considered, like it does not need information like a name that doesn't really contribute anything to a ranking or to a, to a result. so, from like a purely, even just from a

purely model training perspective, it would actually harm the model of including that information because it would try to draw conclusions that don't actually help in the ranking. so both from a technical perspective, it actually doesn't make sense for us to include that information. And then from a practical perspective, there's also kind of a very sound logic behind not including it because we know as recruiters or people having gone through that flow,

We don't need that information to make initial decisions.

Marty (25:05.253)
Right, right. Now, in terms of, like, where how, how widely adopted is people GPT, like as a company, where, where do you stand sort of in terms of the market and what you're going after these days?

David Paffenholz (25:22.446)
Yep. So we just passed 500 customers, which was an exciting milestone. We started off with a lot of individual or small recruiting teams. And so be that individual agency recruiters or freelance recruiters or recruiting teams that were at small companies where it was maybe one or two users. And so that was very much for the first three or four months where a lot of our usage was coming from. But then pretty quickly, we started getting more more pull from larger teams. And so we had to make sure that the

Marty (25:27.217)
That's awesome.

David Paffenholz (25:52.11)
One, the platform was ready for that. had the features to be able to handle that. And then two also go after that market. so every month we onboard kind of a new and larger teams into the platform. In some cases, 30 seats, 40 seats, et cetera. And so the usage there and the requirements end up being quite different. But for us, it's really important that we stick with our kind of product -led growth. And so anyone can sign up. Anyone can try it out. It's free to try it out.

Marty (26:09.989)
Mm.

David Paffenholz (26:20.192)
And it takes just a second to create an account. And that's really important to us. I think that's where lot of our growth ended up coming from. And also some of our virality where people can try it out and tell others about it too.

Marty (26:31.377)
Yeah, that's it's that impressive of our product is once you get your hands on it, people start talking about it, which is kind of like a nice, it's a nice sweet spot to land in. by nature, I imagine of getting 500 customers, you're probably no longer just North America. You've got customers and other geographies and you've talked a bit about that, but I'm curious in terms of, is there like a challenge that you didn't necessarily anticipate with expanding?

globally and how did you address that?

David Paffenholz (27:04.448)
Yeah, so many challenges. The piece that now seemed obvious to me that I didn't think about is like localization of software. Like in theory, every button has to be translated into different languages if a software is properly localized. And so we thought, hey, know, we have data globally, our profiles can be found globally, and the searches work very well globally. But we didn't think about the fact that all of those searches

Marty (27:06.673)
The pain point.

David Paffenholz (27:34.072)
The searches themselves could be in different languages, but the actual platform itself, all the buttons you see are still going to be in English. And so we kept on kind of going through small pain points all the way from how people purchase software in different countries to how tax law works in different countries. And so even the parts that were not as exciting to go through, we've gone through now. And I think there's...

been a lot of growth in some international markets for us. So especially in Europe, we keep seeing more and more people adopt the platform. And also in Asia, we've seen some good growth more recently. There's also a lot of nuance in how the recruiting markets work quite differently. And so for example, in Japan, the coverage of social profiles or of LinkedIn profiles is a lot lower than it is in other countries. In China, it's a very different recruiting system in general.

in some European countries that kind of outbound emailing is less common than it might be in North America. And so there's a lot of nuances and lot of adaptations we've made to the platform because of that as well.

Marty (28:40.625)
I'm super curious. Do you have any other interesting insights from other geographies? I've done a bit of international recruitment, but I'm curious to get your high level take on what some interesting aspects of recruiting in different geos is like.

David Paffenholz (29:03.318)
Yeah, in general, think one thing that surprised me or that I learned is that the kind of the nature of outbound recruiting or of sourcing is much, much more prevalent in North America than it is in most other geographies. And so in many cases internationally, that's only done by agencies or recruiting firms where corporate teams don't do it as much or if so, it's a very new initiative. so especially in parts of the European market, we've seen companies

create their first sourcing teams or create their first active outbound recruiting teams, which was surprising to me because I would have thought that it's kind of, they already existed or that it looked a little bit more like the North American market. And so I think that was one piece that we learned quickly and I wouldn't have expected before.

Marty (29:49.585)
And in terms of, you know, by nature of being an AI driven organization, it's pretty easy for tech companies to adopt and embrace something like people GPT. I'm curious. I imagine by nature of having 500 plus customers, you've gotten some interest from clients outside of sort of like tech. And if there's any interesting insights that you've

learn from from working with clients that are not necessarily in tech and because AI adoption in general, right is is easier in a tech company because you're using tech all day long. But when you aren't a tech company, it you know, AI is a much bigger monster or beast to sort of like embrace at the door, so to speak. So curious what what insights you have there.

David Paffenholz (30:42.636)
Yep. So initially the majority of our customers were in tech today. That's no longer the case. The majority of our customers are not in tech. And I think the kind of excitement for AI or that the potential for adopting it is strong across the board. we see almost in every industry that I can think of, we've gotten interest or currently working with customers that want to bring AI solutions into their companies or into their teams. so ranging from kind of

Marty (30:48.913)
Hello.

Marty (30:55.043)
is strong.

David Paffenholz (31:11.182)
Fortune 500 companies that traditionally might move a bit slower in terms of implementing technologies or looking at new technologies. We've seen dedicated kind of AI innovation teams with them, dedicated focuses on bringing in new AI technology and a lot of enthusiasm for doing that within the HR or specifically the recruiting space as well. And so for us, that's been amazing because we get excitement from our customers or there's a lot of excitement of bringing on board new platforms. And that's something that we really want to lean into too.

Marty (31:41.901)
Awesome. Terms of what you're most excited about for.

for juice box slash people GPT like what's in the, is there anything in the oven that you're like, okay, I can't wait to taste this or anything in the freezer rather.

David Paffenholz (32:01.934)
I think a lot of the potential in the coming months or even the coming years will come from being able to adopt advancements in AI more generally or in large language models and bringing those into applications. And so that's how I view a lot of the work that we do is our responsibility or our job is to see, this is the current state of the art in terms of...

what's available on the AI front and the models. Now, how can we actually apply those to solve real recruiting challenges? And so right now, kind of the forefront of what we're working on is what we call our autopilot. It essentially allows the user to provide a list of profiles that can be from a search, it can be from kind of setting an initial set of filters, then having the AI evaluate them one by one in terms of your fit to the initial prompt. And so it'll

provide you with reasoning, it'll provide you with a percentage match score and more, and provide that kind of really nice stack rank list output that's designed to kind of avoid having to click through profiles at all. And so instead of reviewing profiles individually, we get presented with a list and we're able to take them from there and ideally reach out to them directly. And so getting that process to be as good as it possibly can be is something that is our goal and that I think we'll be able to get closer and closer to over the next three months, six months and more.

Marty (33:25.307)
Is it fair to say that people GBT

kind of empowers a sorcerer to become a recruiter or vice versa empowers a recruiter to become a better source. Like by nature of being a recruiter, you need to be a sorcerer. Is it, is it fair to say that? Or is it just another tool that you kind of add on or because there is a, there's a level of just like having used it myself, like there's a level of disruption or so to speak. like it might seep into the fears that a sorcerer might have if they've never used an AI powered tool.

to say like, you know, this, you know, this is going to take my job, my job, either going to change drastically or evolve drastically or no longer be there. I'm curious to get your take from how this impacts those types of careers.

David Paffenholz (34:14.678)
Yeah, I think the way you framed it of kind of a recruiter being able to also do what a sourcer does and a sourcer also being able to do what a recruiter does definitely goes in the direction of where I see the future as well. Technology is able to help make us more efficient at processes, but it also means that there will be people specialized at using that technology and being able to use that technology as effectively as possible can make an outsized difference. so kind of a comparable piece is if

when say LinkedIn recruiter came along and suddenly there was a database of all profiles that was easily accessible, that was actually searchable and filterable. Now one could think that that would kind of remove a lot of the friction that existed previously and it did, but what that resulted in is actually much more people using it because suddenly more people could be doing recruiting where it didn't make sense economically to do that before. And so similarly our vision is there's going to be a lot more potential for actively searching for the right candidates, including in positions where that's

currently not being done and being able to do that at greater speed and at greater effectiveness. And so I think it'll actually ultimately result in there being more recruiting going on, more sourcing going on, and that also increasing the human component.

Marty (35:28.795)
This feels like a controversial question to me, but I'm going to ask it anyways. Does this like I'm thinking in terms of budget, right? If I'm, you know, my, my past, I've been a head of talent. I've had eight, eight or nine recruiters reporting into me. I looked at my LinkedIn bill at the end of the year, or I'm doing budgets and like, right. It's an expensive tool and they have every right to be an expensive tool to a certain extent.

Is it, can you get away with migrating as a talent acquisition leader, migrating from like LinkedIn recruiter, even LinkedIn recruiter light and using people GPT, or do you miss out on like, is it, is it meant to be? And the reason I say this is like a controversial question. And the reason I think it's a controversial question, you can correct me if I'm very wrong, is

because I imagine like by nature of, you know, people GPT searching LinkedIn, there's, there's a partnership or some form of connection to LinkedIn that they're very aware that you're going through profiles.

David Paffenholz (36:41.942)
Yeah, it's a good question. So I think we have seen customers kind of stop using their LinkedIn and purely relying on people GPT for their search and outreach. I'd say the main thing that I kind of use as the deciding factor is how much of your searches are for people looking for people that are open to work. And if that's a frequent use case, then we're probably not going to be the best solution because we don't have that as a filter. And if that's not the case, if the majority of searches are not looking for people that are open to work.

we may be a better solution because I'm confident we can provide you with the right matches more efficiently. That also means that we're able to showcase data that or profile data based on not only LinkedIn profiles, but also data that's available beyond that. And so a big focus of what we do too is how much other data can we find about an individual? What else is indicative of their skills, of their location, of their profile that we can use within our profiles? And so I'd say,

those two pieces, like one, kind of how big is the focus on open to work and two, how much other data can we gather would be what influences my recommendation of whether it makes sense to move away from LinkedIn or not.

Marty (37:51.793)
Awesome. No, thank you so much for your time. I had a blast learning about people, GPT, learning about you and your track record. I love the tool. Thank you for making it. It's been really cool to learn about.

David Paffenholz (38:09.6)
Awesome. Thank you so much for having me on. This was really fun.

Marty (38:12.603)
Yeah, no, anytime. I'm glad you have done it and yeah, well, I'll reach out if I see anything crazy going on or if you go viral again and maybe we need to need to do a revisit and see what's up again. So thanks again. Appreciate it.

David Paffenholz (38:21.847)
Thank

David Paffenholz (38:27.97)
Please do. Thank you. Have a good one.