Steve Bartel Live on Shally's Alley

Hey everybody, happy July 11th! It is Friday and you are live on Shelly's Alley. Welcome back! Yeah, peace. If I do that, I think I get some kind of weird effect.

Yeah.

Oh yeah. So here we go, super excited. My buddy Steve Bartell is our guest today. Steve started a tiny little company you probably never heard of called Gem. It's like a little diamond. Yeah, you probably never heard of it. Yeah, never have. Changed names. It used to be called something else, but it's all good. So thanks for joining us today, man. Quick background for those of you that don't know. So Steve is actually the founder, co-founder, and CEO of Gem. CEO of Gem and frankly one of the sharpest minds that reshaped modern recruiting scaling of teams. Ted has taken Gem from, you know, a startup all the way through to a company that is essentially creating some of the innovations that others are following. And he's got a bit of a history. Why don't you introduce yourself and tell us a little bit Real quick about you before we jump into the questions, Steve.

Yeah, that sounds great. Before I do that, super cool theme music. Do I get to customize that next time? Like, do I get my own theme song? That, that was, uh, that was a fun intro.

Yeah, I can get you a theme song. This new— this thing now is, um, now does generate— generates music. So believe it or not, it's computer-generated music. So I can do like chill What? Yeah, check this out. Computer generated.

Yeah, it's so weird.

It's got like, uh, there's some genres that I've never even heard, like Minimal House. What is Minimal House? I have no idea what that is. Let's play that. Minimal House. Oh. Talk about like next-gen news, right?

Wait, that's, that's wild. And so that's all AI generated?

It's, yeah, so no digital rights or anything. Yeah.

Is that through the streaming platform you're using or you have a separate tool for that?

Oh wow, it's this streaming platform. I do have Stream Deck and I use that a lot, but, um, I like this kind of other stuff in here. But most of the time I want to focus on my guest, so I end up not playing music. Um, it's built in, so it's pretty sweet.

Yeah, that's so cool. I got to get my team on this platform. This is so cool. I mean, the future is— it's going to be interesting with AI. I'm sure we'll probably spend some time there today, but it's— yeah, great to be here.

It's kind of a beaten up topic, but I do have a question about that.

Yeah.

Why don't you tell us a little bit about like you were— you're a recovering engineer. Yes. You used to be at like you got your formal, I guess, engineering leadership at Dropbox, but you were at Facebook, you were at Blizzard.

Yep.

And then you decided— I mean, get that— my first question would be, why? You had a good thing going with Dropbox, and you're like, I'm just gonna go and start something, you know, and deal with recruiters. I mean, what happened there?

What went wrong? Well, first of all, I love it. Oh my gosh. Like, one of the things that drew me to the space was I just— I loved working with our recruiting team at Dropbox. I feel like as an industry, like recruiters are— and I myself have done a ton of recruiting and I just feel like any of us who got into the business, like we're people, people, people, like we do it because we care about candidates. We care about, you know, bringing in the right people. And, um, a lot of us care about the impact that that has on, you know, each person that gets hired, right? That's like life-changing. And so, you know, I think though, wow, you've had any success, that's—

I mean, I put that down as one of the top 3 motivators. You know, people don't usually get into recruiting intentionally. It's unusual. It does happen. I used to teach a class at Temple University where we were taking people from their HR MBAs directly into recruiting, but it's unusual. So most of the time we fall into it, but it's one of the top 3 reasons we stay. Like, it's not one of the reasons we get in, but once we're in, it's one of the top 3 reasons why we stay in recruiting because of that win feeling, especially when you make a hire, make a placement. Where the person and the job would never have met under any other circumstance. Like, you're, you're literally the intel that made that happen.

Totally. And, you know, I think you're right, actually, now that I think about it. Um, I didn't have a taste of that until I started doing a ton of recruiting myself at Dropbox, and we can get into that in a sec. But, you know, I think the thing I've come to realize, you know, now that I've been doing my own hiring for, you know, 15 years, but also just been in the industry for so long, is that initial rush of making a hire, the even more meaningful thing is like seeing somebody like, you know, grow in their careers and get promoted like year after year after year and seeing that like life-changing opportunity like come to fruition. You know, the post-hire as we, you know, think about the impact that we have like that is profound. But yeah, no, I think like Dropbox was actually As somebody who was on the engineering side, um, I was thrust into managing engineering teams less than a year out of school.

Yeah, you went to MIT and, um, with a computer science degree, right? You graduated.

That's right.

And so, I mean, you've— you haven't even been in the workforce for all that long, considering.

So yeah, it's pretty amazing, like 15 years, right? Yeah. And I've only had 2 jobs. I've had, uh, well, I, I interned at Facebook and at Blizzard, and those were incredible, you know, places to, to work. And, you know, Blizzard was a little bit of a passion internship for me because I grew up playing all the games. Like StarCraft II was actually the game I got to work on.

Oh nice.

That was a dream internship. And not only that, I got to build the AI for StarCraft II. So whenever like a unit would cast a spell, it would run my little AI to decide you know, is it— is now a good time? Or, you know, who should we target? And it was just such a, such a cool first industry experience and kind of, kind of the dream internship. But yeah, Facebook was incredible too. But I think like Dropbox, I joined when it was 25 people and we just started hypergrowth right around the time I joined. It was this inflection point. For a long time, actually, the, the growth of headcount had been relatively flat. Um, the company was 3+ years and it was 25 people. Um, but the business had been taking off. Uh, and so I think Drew realized, um, wow, we, we've got something. We've got like lightning in a bottle here. We just need to like hire as quickly as we can to go after the opportunity. And to his credit, and also to Aditya's credit, the VP of engineering, that he brought on board from Facebook, actually, they both instilled in all of us that our number one priority, the single biggest way any of us could have an impact was hiring, right? Because like, and especially as like managers, you know, I was thrust into—

right there.

What's that?

That's a Steve Ballmer move. I used to work at Microsoft. He was big. And I mean, he was very much like our new hires, our hires are our lifeblood if we don't hire the right people. And he put a lot of emphasis on recruiting. So yeah, totally.

I mean, I think that's smart. I think that's really smart for every CEO.

That's how you got into recruiting then.

It is. Yeah. And so I was like an eng manager straight out of school, like 9 months into after graduating, I was managing an engineering team because we didn't, we were growing so quickly. I was the most senior person in the area and like we needed a lot of help. And so for many years there, I just really leaned into recruiting because the single biggest way I could increase my impact, it wasn't like going and writing more code. It was how do I, you know, bring more great people into the business? And so I got to do everything. I, you know, obviously did a lot of the things that most engineers do. Tons of interviews, referrals, you know, working with my own team on like referral jam sessions, going on campus, giving talks, all that good stuff. but actually I got to like really lean in in some other cool areas. So one year I actually revamped our entire referrals program with our head of rec ops. Like her and I came back.

I did that at my last company as well from scratch.

You did?

Yep.

Oh my God. Yeah. Well, I mean, that is so cool.

It was one of those like people were just being bombarded with like, hey, referral bonus, and they wasn't working, you know? So yeah, completely revamped it. Yeah.

Yeah, I mean, that, that we could go, we could spend a whole session on that, but like, that was our life for the next one. Yeah.

Platform, does it?

No. So yeah, let's, let's get back on top. No, I'm just kidding. I mean, I love to start shooting the breeze on, on anything recruiting related. So yeah, but I revamped our referrals program. I actually was my own full cycle recruiter to hire a few data scientists. This is a fun one. Back in 2013, did you ever recruit for data science?

Oh yeah. Yeah.

Oh boy. Do you remember 2012, 2013 era of data science before it was like a real thing?

It wasn't even really like a thing yet, right? Yeah. They had— it was big data and people were calling themselves data scientists who were actually analysts.

I know.

It really— like, when you say data science, Do you mean true? You're actually writing statistical analysis programs?

Yes.

R or something similar? Or are you using a spreadsheet to make a chart? Because that's not—

or just running SQL queries. Yeah, no, I know it was so confusing. It was so nascent. The field was just developing and we were like going after PhDs with quant-like backgrounds, people that could help us analyze the hundreds of millions of user data.

Big data.

Billions and billions. Yeah, big, big data. And so I went and hired our first few data scientists. I still remember I was like, hey, Aditya, I think it would be really impactful if we built a data science team. He's like, yeah, go do it. And then I went to our recruiting team and I was like, hey, like, I just got the green light to go start a data science team. And they're like, what's data science? Yeah, well, but also they're like, yeah, go, go do it. Go do it. Like, we have this, we have this tiered role system. And, and, uh, it was really interesting because we were growing so fast, even recruiting for recruiters, um, we were struggling to keep up because we were tripling headcount every year for the first 4 years I was there. And, um, yeah, they're like, look, like, we've got all these other higher priority roles. Like, we'll have somebody that can help with scheduling, but you have to do all of your own, um, you know, sourcing and even like full cycle seeing them through the process. And so You know, I got access to our ATS. Um, I was like, you know, the recruiter on the job. Um, and then— and you still have a career.

That's amazing.

I know.

Well, yeah, I would have been pulling— the word ATS literally makes my— the back of my eyeballs itch so much that I want to pull them out.

You got to get on Gem ATS, my friend.

It's, uh, But yeah, the old school ones and what— yeah, I'm actually doing a whole comparison chart right now anyway.

So, oh my God.

And you're like, wow, this is crazy. Why is it all so bad?

Why is it so bad? And actually the hardest part of it was like we posted a job. I went and gave talks down at Stanford to the PhD grads. Like, you know, we— they took you at Stanford.

They throw a bunch of rotten eggs at you because you came from MIT. I mean, come on.

I know. And I was an undergrad from MIT. Who had no background in statistics or data science. And I was going and—

What's this guy doing here?

—to, yeah, like 300 Stanford PhDs with quant backgrounds and highly technical backgrounds. And yeah, they didn't laugh me out of the room. Actually, that's where I met the first person that we hired. Was down there, but also like it wasn't enough. Like we needed to hire a lot of data scientists and this was a nascent field. And so I talked to our recruiting team, they're like, oh yeah, yeah, you gotta go source. And so I was like, well, what's that?

Just gotta go do it. Yeah. Like, yeah, what's that?

So I got on LinkedIn and I started sourcing people and like sending them an InMail. And I was like, I'm not, nobody's responding to me. I wanna ask my recruiting team and talk to like one of our, most experienced sourcers, this guy DJ back then, he's like, oh yeah, yeah, you gotta use email 'cause like these people don't have their LinkedIn notifications on. I'm like, well, how do I find that? He's like, oh, like go use this tool or actually check these 5 tools to find email. And then I started using email and I was like, I'm still not getting the responses I need. Oh, he's like, you gotta follow up. I'm like, oh man, all right, I gotta track this stuff in a spreadsheet and, you know, Let's see who I've followed up with and sent all these manual follow-ups. And oh my gosh, I mean, that whole experience felt really broken. Yeah. And so, but I, I think like all that, I mean, look, I even back then acquired two companies that I sourced, one of them for like acquihire, like it would, that's how important recruiting was. And like these were all the different phases.

My friend Maureen talks about that all the time, like sourcing companies to Sourcing companies to hire the company by buying the company.

Yeah, exactly. Yeah. And I was actually doing it through my network. I was just reaching out to smart folks, you know, seeing who might be looking for a home. And we hired some, some great people that way, but we were just trying everything. And so I feel like Dropbox was where I really got into like sharpen my, my skills and fast forward 5, 6 years later, I mean, it was really just 1, 2 years after Dropbox, um, I was like, man, like recruiting was really tough. Like maybe, maybe we should go start something in this space. And there was all this amazing technology that was being built for like sales and marketing teams. Mm-hmm. Yeah.

And there's always been that story of like, yeah, I mean, I've been doing this for 30 years and I, I've multiple times tried to like adapt technology that was being used from, you know, sales, affiliate referral programs, multi-level marketing, all these other technologies that were out there. Um, and sometimes some of that stuff works, but eventually there's a just a deep chasm that appears between marketing and sales, or digital marketing, and all of that stuff, all, all of the non-recruiting related and recruiting. There's this big chasm that forms And it's impossible to get over unless you fix some of the things that are, that are, that are, that don't work in marketing. Because for example, you have the two-way relationship, right? Marketing is a one-way relationship. Sales is two-way, but with sales you have one person who's the decision maker. And in recruiting you have three. It's, you know, so there's these big breaks.

And, and even more than that, like more breaks, like, you know, I think in sales, um, you care a lot less about long-term relationships because your stakeholder at a company switches jobs and now they're no longer relevant to your account. Whereas in recruiting, like we get to know people and we track them through their entire career. And so like long-term, well, yeah, I mean, we're supposed to, if we have the right tech to help with that, uh, hint, hint, you know, philosophy.

Sometimes it's very transactional too, but yeah.

No, I guess that's— you're right. I do think a lot of modern tech companies and even large enterprises are thinking that way though, even if a lot of, you know, other, you know, recruiters might be a little more transactional. But I firmly believed in— actually, at Dropbox, we believed it was all about long-term relationships. We had these spreadsheets there that we used to call our list of gems. Uh-huh. Ah, that's where it came from, right? Yeah, yeah, exactly, our gems. And, and we, we literally called this as like Steve's Gems, uh, Shelly's Gems. And these were the like 10, 25, maybe 50 tops people that like we went to school with or that we had worked with in previous internships or previous jobs where—

well, you would hire if you could hire them in an instant, right?

Of course we're going to put them through the interview process, but like It was our shortlist, our gems. And so, yeah, I mean, I think a lot of that was the inspiration. That doesn't fit in an ATS.

There is no, you know, the ATS, the, the, the birth of the ATS. First of all, if I had a time machine, you know, people ask me that question, what would you do if you had a time machine and you can maybe wave your magic wand and, you know, go back in time or whatever? The first thing I'd do is I would abort the ATS. Like I would prevent it from existing at all. It's the absolute worst thing that's ever happened to our industry. It literally exists only to mitigate risk. That's it. And it doesn't track applicants. The applicant tracking system tracks requisitions.

I know. Exactly. Exactly. And actually many of the large, like legacy enterprise ATSs still don't have a person record. Nope. All they have is an application record. And so talk about transactional recruiting. Right, a rec and like an applicant. And, um, like, how do you, how do you try, how do you track relationships over time when you have—

yeah, people ask me about that, dude. I get this question a lot. Why can't we use an ATS as a CRM? I'm like, it's a 90-degree— they don't, they don't intersect. It's because of what you just said. The other question that I get too is why can't we use Salesforce or something like that as a CRM. I will tell you this, I don't know if you knew this, but I was the first person— my friend Eric Jacob and I were the first to do a recruiting CRM using Salesforce before Salesforce was, you know, we're talking about 2003.

Wow, that's so ahead of the curve.

And yeah, we've been— because I needed something to do recruiting CRM and the ATS didn't cut it, and none of the sales CRMs out there— I mean, I had tried Infopath, if you even remember what that is, Dynamics, a whole bunch of them, and it didn't work. And so Salesforce was the one that had the most DIY stuff where you can go in and, you know, and, and like create things. And so I made it work. It was absolutely off— not because it was Salesforce, it was awful because it was some guy that doesn't know anything about, you know, databases. Like, I was barely an engineer inventing a recruiting CRM back in 2003. But people still ask me, why can't we use a CRM for sales? Like, even that doesn't, it doesn't, imagine if you had a sales CRM where instead of selling to a person your company's product, you're selling them every company's product. That would be like, Try that. Yeah. It doesn't work. You can't.

Yeah. No, you're totally right. Cause like every job is different. Every rack is different. Yeah. No, there's a lot, basically there's a lot of differences. I mean, I think like the, the data model needs to be different and some of the differences we were talking about in recruiting about being able to track long-term relationships over time, because like as soon as your stakeholder changes, you don't care about that person. In sales, there's this whole, ecosystem around data enrichment, right? Which, um, you upload these big lists of contacts and work emails, but like in recruiting, you want to use personal email. That's evergreen. It doesn't change over time. Um, and you don't have recruiting ops teams that can go data enrich your, your database. Like, so everybody's like searching for these people and trying to look up their emails. And so, yeah, I mean, there's like a myriad of reasons why you need something purpose-built for recruiting, but the, the The fascinating thing is like, you know, Nick and I, my co-founder, we were tech people, we were engineers, but we just saw our friends falling into these traps of like going and starting companies like with technology in search of a problem without first validating the problem that was real.

In search of a problem is about how I would describe most of them.

Yeah, most, most, I'd actually, you look at most recruiting startups, like our startups in the recruiting space, and a lot of that's just like, tech in search of a problem because, you know, some engineer realized that recruiting's hard but doesn't actually have like the full cycle recruiting experience or like hasn't been in the life of a recruiter to know what the problems are. But still, we were like super paranoid about that. So we were like, we got to go talk to 50 other companies that aren't Dropbox and Facebook and validate that this is a problem. Because like, we didn't want to go build something for just 2 companies in the world. What if nobody else thought this way? And we did it. And we, we talked to 50 other companies before we wrote a single line of code. And we found something really remarkable. Well, first of all, everybody had this problem. Everybody was thinking about recruiting more like sales and marketing. But the cool thing was, just like you back in 2003, 10% of the most cutting-edge recruiting teams were using sales tools, right? Because for their sourcing, closest thing you could—

yeah, yeah, it was. And, you know, it's not quite there, just it was closer than everything else. I ended up at Microsoft making my own later on after the Salesforce one, because once I went to— left Google— once I left Google to go to Microsoft and, um, you know, wanted to create a CRM there, obviously I couldn't do Salesforce given that, you know, you kind of got to eat your own dog food. So, um, I ended up creating it from scratch with InfoPath, which was later absorbed into Dynamics. And to this day, I think— I'm not sure if anybody's ever been able to do this, but it was the largest database of computer science graduates with more than 5 years experience in the world. It had 250,000 people in it. Wow. Or 2005, in InfoPath. I love that it worked because it was a homegrown CRM for recruiting, you know, and then eventually InfoPath got sunset and whatever.

Tragic.

Um, yeah, I mean, I feel about AI though because I know we— you wanted to touch on that and I did have a question. You've built a platform and, you know, you were in AI before it was— you were in data science before it was data science, you were in AI before it was, you know, cool, right? Like, I mean, I was in robotic process automation for years doing all these RPA with Microsoft. You remember? Cool. And stuff like that. Yeah, that was— that's what AI is now, right? It's just the evolution of that. But anyway, you have this AI that can remember conversations from 18 months ago, and I can't even remember where my keys are. Um, where do you— where's the line that you draw between human memory and machine memory in recruiting? More specifically, are we too quick to outsource our intuition to tech?

That's a really good question. So I actually think tech has already augmented our memory big time. I mean, even just think about like a digital calendar.

He doesn't know where my keys are.

Unless you have a Tile or an AirTag.

That's what I did. I got an AirTag.

Yeah, exactly right. And so now tech knows where your keys are. No, but seriously, like, tech has been outsourcing or like augmenting our memory for a long time. Like, augmenting.

Yes. Yes. That's my question. Where do you draw the line?

You know, so here's my vision for how this stuff works. So first of all, what you're referring to is we have some really amazing AI Um, all right. So my unique perspective on AI is that it can't just be a bolt-on point solution. That's right. Like this is Jem's perspective on the market, that your AI, it's gotta be native. It's gotta be native and it's gotta be native and it has to be a part of like an all-in-one recruiting platform. And here's why the key thing that nobody realizes in the space, there's all these AI upstarts, but nobody, nobody really realizes, even most buyers don't realize the hard part about AI is no longer the algorithm. OpenAI, Anthropic, like they've all solved that and it's— we can buy that. Any of us can go see it for ourselves. Like with Claude and ChatGPT, it's so good and it's getting better like every month. And so that's not the hard part. That used to be the hard part to build effective AI. You used to have to hire like, you know, whole teams of machine learning experts, deep learning experts, and train it on like—

it wasn't really— it's really still not AI. It's still, you know, language models, but the leap was the ability to be— to bring it to every day. So it's kind of like I refer to it as, yeah, right? You can't build software that uses the internet as an add-on. The software has to be built internet native. And so that's what's happening is it has to be built AI native. It's the same idea.

It does. It does. Yeah. So, so all of the software needs to be built AI native. And here's one of the key reasons why. Outside of just like wanting fewer tools, everybody wants fewer tools, but like the key reason why is the hard part about AI is what context, what data does it have access to, to make smart decisions, to make smart recommendations. And like a really simple example is like you wouldn't want an AI sourcing bot going out there and just burning down and spamming the market without any context about who these people are, especially if you're like a company with more than 10 employees. right? Like any reasonable company that has like done some recruiting, like you, you don't want it to like go shoot itself in the foot. So you wouldn't want AI reaching out to somebody that your recruiting team already talked to 2 months ago.

Okay, so this is where I'm going with this is I want to know where that line is. Yes. I'm writing a research paper now about augmentation versus automation. And from the data that Claude provided and a few other places, it's pretty obvious that most of the people who know what they're doing that are using these tools, they're in Claude like me every day, and they're in, you know, ChatGPT every day. 80, 90% of what they're doing is augmentative. 100% of it is this agentic. And so my point, Steve, is this whole agentic thing I don't think we're there, and I don't think we should be there.

No, that 100%. No, you're speaking my language. Look, a lot of companies are calling their AI agentic. Like, maybe Jamf will even brand our AI as agentic. It's way more agentic than, like, most, but, like— From a marketing point of view, I get it.

But really, it's 20 people in Ohio that are doing all the manual stuff that the AI can't do. They're lying. It's the same. Yeah, no, totally.

Totally. And that's been the state of AI forever, right? Like anyone— Exactly. So AI is not replacing humans, it's augmenting humans. And we're seeing this across every industry. And I'm happy for that.

I just don't like it when people say, well, I'm not going to buy your product because it doesn't have agentic.

Like, well, I mean, that's why like everybody's marketing their AI as agentic. But Jem's approach is is different. Like, we are, um, 100% in the copilot augmentation camp, and we believe that's the future. That's the way to make AI that actually works. And anybody that's selling AI that can fully replace humans is selling vaporware.

And we've seen this play out. It's not that replacing humans, because we know that's not happening. It's not happening. Judgment line where we get into, like, I mean, I have a hard line on ethics and it's a pretty hard line I've been speaking about openly, but I want to hear from a tech person that has a product in the space. Where's your human judgment line where you know it would be best if we don't let AI make those decisions? Yeah, great question. So Help us with everything else. Million ways to help us, but not that.

Totally. Here's, here's a really good example. In my mind, like an AI agent that helps you with your inbound applicants, it shouldn't actually be rejecting candidates. Yeah, it should be speeding you up as a recruiter. It should be like, so it, what it should be doing is like giving every candidate an equal shot regardless of when they applied. Because today the state of the world is the default is you post a job, you get thousands of applicants, and a recruiter is like, oh my God, I got— first of all, I'm going to go turn that job off. I'm going to take it off my career site. And now the best candidate might not even get a shot at the role. Right. But then the other thing that happens is like a vast majority of the people that are hired are like in the first 100, 200 people who apply. And that's not good either because if, you know, you find your dream job and if you're the best fit for the role, you're not getting a fair shot if you're applicant number 2,000. So, you know, 20 plus, maybe closer to 30% of our customers get thousands of applicants for a single role. And so without AI to help like, you know, sort through that and elevate some of the like shortlists to the top, It's not giving everybody a fair shot. And so that's where I believe AI can really help. But what it shouldn't do is go reject people for you because now you're running into ethics issues about like—

That's my line right there. Now there's ranking and scoring and sorting and weighing and all that stuff. But the moment it is making a decision. Exactly. Yeah. So it's like, let me put it in terms that I think a lot of people who are listening would be able to relate to here. In my opinion, after being in the business for a very, very long time and helping more than 500 companies build their sourcing teams, I can tell you without a doubt that the most value we offer sourcing and recruiters that source, the biggest value that we offer is decision support. We are invaluable and irreplaceable when it comes to providing leaders or hiring managers with the information they need to make a decision. And so that's why when people say we need to measure the sourcers based on their number of hires, I say stop. Because if as a sourcer, put yourself in your hiring manager's shoes when you were back at Dropbox. Let's say that you had an opening and you only had one finalist, one candidate. That's it. You have no choice. There are zero choices. And I, as the sorcerer, come in and I present you with a second candidate that is pretty good. You even consider them and interview them, but ultimately you decide to go with your original candidate, your gem, right? However, I now enabled you to make a decision that you previously were unable to make. In the book of let's count hires, I did not do my job. Yeah, yeah, that's not— yeah, no, I very much did my job.

I gave you way more to answer. Yeah, decision support.

So I see AI exactly the same way. It's providing me with decision support. It's not eliminating anybody, but it is helping me surface. Maybe at another time I'll tell you about the rubric and things that I'm building to hire the country's top scientists, but I'm doing it to include people that were previously left out, just like you just said, because it's decision support for me. Then I take my top 20 and send them to Steve, and Steve measures them against his gems and makes the decision. But you're the hiring manager. I'm never going to make the hiring decision unless I'm on the team, obviously.

Yeah, no, I totally, I totally agree. And, and here's, here's the vision for how personalization could work at scale. Using candidate relationships and history, which by the way, like you only get the very best like candidate relationships and history context if you have a complete source of truth, which is unlocked by having like an all-in-one recruiting platform where all the pieces are fully integrated. So like that's going back to like, that's, that's our vision for AI needs to be built in an AI-first way into an all-in-one recruiting platform. And we're the only ones that are thinking that way and doing that. Um, the vision here is 'Cause that way you get like full, you get complete data on every person that applies to you.

Yeah, yeah, somebody you forgot about because you're human 18 months ago applied for the job and it's perfect, but they didn't reapply because they got disappointed last time and they never really came back even though they're still interested. If you called them, but you don't remember to call them because you're looking at 2,000 applicants, the AI will say, yo, tap, tap, tap. I'm trying to tap the screen. Tap, tap. Hey, Shelly, remember the clip? Microsoft clip. Tap, tap, tap. I see you're looking for a candidate. You should consider this. I'm literally imitating the clip now, but you know what I'm saying.

No, exactly. That's right. And like the AI should know that somebody interviewed a year and a half ago. They should know that you met them on campus like 4 years ago. It should know that somebody attended one of your recruiting events last year. It should know that like Shelly on the recruiting team, like, you know, talked to this person 9 months ago. It should even be able to like understand little nuggets from those conversations, from the interview feedback, and like work those things in to the personalized outreach. And so that's like where I think this stuff is headed. Hyper-personalized re-engagement based on the full relationship history and context. And like, that's been like the biggest treasure that every company is sitting on top of. And doesn't know how to unlock. Here's some stats for you. Okay. This is gonna blow your mind. For a, you know, downmarket company, few hundred people, already 30 to 50% of the people they hire are already in their CRM or their ATS from a, a previ— previous touchpoint, right? And so that's, but, but everybody's going and spending more recruitment ads to get them into the process or like sourcing them manually on LinkedIn instead of just tapping into their database. For an enterprise customer, thousands of employees, 10,000s of employees, 50 to 70% of the people they hire are already in their system. Yeah. But like nobody's able to unlock those people. And so they're like going and redoing all the work. It's almost as if in recruiting we suffer from amnesia. And so, but that's because of the way that it would be.

There's some regulatory things that, that, that apply, and people that are not deep into the business aren't as aware of the regulatory, um, things that prevent us from having historical knowledge. The amnesia is sometimes, uh, mandated.

It is mandated for highly regulated companies, especially in EMEA, that are really large, that, uh, are a little bit too risk-averse. Sorry, that's my hot take. Yeah.

Memory correctly and there's a way to use memory incorrectly. And there is. What's missing is the context, right? Because— Exactly. I can tell you that, for example, every time I go to the gym, I'm sore for 3 days. So the AI can interpret that as never go to the gym so you're not going to be sore. Yeah. Or it can interpret it as— No, totally. What are you doing that makes you sore and what can we do so you stop getting sore when you go to the gym because you want to keep going to the gym? It really does matter. And this, again, we'll go back to that decision support.

Yeah, I don't know, 100%. But even the companies that are highly regulated that can't keep all their hiring data, um, they maybe have data retention policies that they have to like, um, delete those records after a certain amount of time. They can still keep historical context without the— some historical context without the like sensitive hiring data that can still help them understand who this person is to them, what kinds of touchpoints they've had, even if the details of like those touchpoints have to—

You know, I learned something really interesting when I was building the global team because I got a— I was actually— I sourced and hired our chief privacy officer who was a former black hat engineer. And I spent a couple of years working with her to help build some of this out globally. And what I found was really interesting that we ended up after going through some, you know, a lot of different variations and analysis, there definitely is a right to be forgotten in this GDPR concept where like I wanna be deleted. However, there are two pieces of information or two types of information that A, cannot be deleted A, or shouldn't ever be deleted, and B, should not be handed to the candidate when they ask for the information. So even with privacy laws, as a company, we have trade secrets, and it doesn't matter what your privacy rights are. There are things that are trade secret to us that we cannot share with you. Yep. Even if you ask for all the information we have about you, we have to redact it and not give that to you. And that's fine because otherwise you're violating our rights for confidentiality of our trade secrets. And then the other thing is there are some things that we need to retain no matter what.

Totally.

Because otherwise, not privacy, but legal, which— so, you know, do you break the law or do you violate privacy? It's like, uh, Yeah, no, it's—

sorry. Yeah. I mean, that's the funny thing about all these privacy— they are in direct conflict with like the legal reason why you need an ATS in the first place. Like going back to your point about like why these ATSs were invented as just like these legal databases for compliance. Like you need to save a lot of those records legally. Yeah. I never— that's cool. I never— I mean, I knew both of those independent data points, but I'd never put them together.

Yeah, it was a very long set of conversations and lots of Slack time with my CPO. But I'll tell you the fact of the matter that she had this black hat background and I do too, secretly, nobody knows. We both arrived at some of these ethical conclusions that were like, okay, there are some things that we simply cannot do. And we're just not going to. We would rather pay the fine, but really that shouldn't exist. That fine shouldn't exist because we, we understand about removing the PII that would be sold to others and used in other ways, but there are some pieces that, you know, you can't do that with. I wanted to ask you, I wanted to ask you about, um, founder blind spots. You said before that hiring should be treated like a product function. And there's been a few leaders in the business world that have said that, which I really align with because it's— we've talked about it. It's not sales. Recruiting is not sales. Recruiting is not marketing and it's not operations. It's a little bit of all of those, but really treating recruiting like a product makes a lot of sense. So if the candidate experience were an app, Um, what feature request do you wish more founders would prioritize but just never do? Hmm.

Yeah, uh, so there's a few things. So I mean, the first, like, the one that I think should be table stakes feature request, if, you know, if we just think about candidate experience as a product is every single applicant. Yeah, yeah, exactly. Yeah, no, and it's— I just think it's like every single applicant should hear back about whether they're entering process or you're rejecting them for a role.

There's that. There's absolutely no reason for them to get— like, there isn't any. There isn't any. There are legal reasons for not saying certain things, you know, um, even if you made the right decision You know, like, for example, I was talking to a candidate, very nice lady yesterday, and I'm hiring for a position that is with a cabinet-level organization. And she very straightforwardly just up and said, so listen, I heard you can't hire Chinese anymore. There was a memo. I don't know if you're aware. There's a whole law about this now. Um, and she's like, I heard you can't hire Chinese people. And I'm like, okay, so here's the thing. We have a lot of Chinese people that work here. We're very international. Uh, yes, we can hire Chinese people. There's nothing wrong with you being Chinese. There's nothing against that. But for— there's a specific job that the job itself requires by law that you be a citizen, naturalized or otherwise. You have to be a citizen or you cannot do this job because of the government, not because of us. But, you know, so we can hire you for 99% of the jobs, and being Chinese has got nothing to do with it. But in this one case, we can't hire you, not because you're Chinese, but because you're not a U.S. citizen. So it doesn't matter if you're Chinese or Colombian like me. Yeah, we can't hire you for this job because of that. And so there are certain things that we put in our documentation that, like, your citizenship status, you know what I'm saying? You know what I'm kind of getting at here? You just— yeah, yeah, so you're right. And, and, and so if our system were to automatically reject her and say, uh, yeah, we unfortunately are not going to be able to consider you for this position because you're Chinese, that's just wrong. You don't do that. Because it's not right. It's not because she's Chinese at all. I know, I know.

And I could totally see AI saying that too, like, and that's where like this stuff gets kind of risky.

If that— yeah, say, thanks for your application, we're very interested in you, you have the background we want, we approached you, we recruited you, we spoke with you, and after speaking with you, we determined that you do not have one of the requirements, which is you must be a citizen. Therefore, we can't proceed with the application, but we would really, truly, honestly like you to apply for other jobs where citizenship is not required. You're actually— nothing to do with it.

No, totally. And I actually— we don't do that. We don't do that. I mean, it's hard to do that at scale. I mean, that's the hard part is like it takes, it takes so much time. Like, you know, recruiters are getting thousands of applicants for a role. Each recruiter is managing 55% more wrecks than they were 3 years ago on average.

Pisses me off, man. Dude, let me put it this way. You could be sitting in the inbox with 2,000 applicants and you could still do the right thing. There's nothing preventing you from doing that. What's happening is they're marking everybody that they didn't look at as other candidates more qualified. And so you can't— oh yeah, letter saying I can't consider you because you weren't qualified, when in fact You didn't consider them. No, I know.

I know. Exactly. And so, but, but here's the thing. So, so recruiters are managing 55% more recs. They're, they have 3x the inbound compared to 3 years ago. And so most recruiters just aren't getting through the full backlog. And so they can't get back to anyone for that exact reason.

System where you can. Here's what you do. You have a category that says I didn't look at you because we closed the rec. Yeah. And that person, those 2,000 applicants get a message saying, totally, we were not able to review your background because we've made an offer and closed the rec or whatever. And this position is now closed. And we did not— your application didn't come in in time. You know, by the time we got your application, we were already over here. So you weren't considered. But we want to encourage you to apply for other jobs. This is not a rejection letter. This is just to inform you that this position has been filled and we couldn't get to you.

Yeah. And so that is one option, but it's still a bad candidate experience. I mean, it's way better than not hearing. Better than that. Way better than not hearing. And every company should be doing that because you don't need anything, any technology to like— it's table— that, that should be table stakes. But from a candidate experience perspective, if the rec takes, you know, 3 months, like, and you don't hear back until the rec was closed and you didn't even get considered. Like, you know, that, that feels pretty shitty. And so, I mean, I think this is where—

buddy, I'm talking about— what's that? I'm not talking about 3 months. I'm talking about when you close that rec and you're not taking any more applicants, it shouldn't be 3 months later. Why are you sitting on 2,000 candidates for 3 months? Oh, it just—

it happens, man. And so, like, because, because, like, recruiters— it's like some recs get 10,000 applicants. I mean, you know, maybe people accepting applicants Maybe people should take it down. Yeah, I look, it's, it's tricky. I, here's the thing though. I do actually think AI is at the point where it can do basic qualification matches based on like what are the minimum required qualifications. That's what you're saying, right?

Matches. Exactly. You fit, but not you shouldn't apply. Right?

Yeah, totally, totally. But like once somebody applies based on their resume and based on the job requirements, you know, AI, I mean, this is the way our AI works is you can tell it here the 5 to 10 like must-have requirements for the role. And AI can actually do a pretty good job of saying, is this person qualified for this requirement? Is it qualified for this requirement, that requirement? And so move those forward. Yeah, exactly. So you can bubble up like applicant number 8,000 that's like Boom, like 10 out of 10 requirements.

Head to the 10,000 candidates that way, and the ones that didn't bubble up were not qualified. And then you can say these were the qualifications that were missed if you have the ability to track that and automate it.

Bingo. And, and so I actually think, I do think AI can, we can get to the point where we can actually give every candidate some actionable feedback about why they weren't a good fit for the role, so long as we're just doing, but that's not basic qualification matching. It's, uh, it, it, so it's, it's leveraging generative AI to— what's that? It's a logic tree.

It's a decision tree. Yeah, exactly.

It's a— no, it is. And, and it's also giving recruiters control because it's like, well, how many requirements, uh, is it okay to like be missing before to be able to put them into the, into the process? But also which requirements can't miss requirements versus the nice to have versus the nice to haves, right? And then how many of the nice to haves can what can they be missing? And so like recruiters still get full control over all that and they can still like review every single person. It's just like really speeds it up and puts the people that match all the requirements—

been able to do this. We've always been able to do this. This just speeds it up. So it's really not— what you're suggesting isn't new. What you're suggesting is the right way to do it, and the technology is there to make it easy, but it's still not being done.

No, the technology is now at the point where it's like 50 times easier to do, to do that.

It's not because recruit—

oh, sorry. Here's, here's why it's not getting done.

We know we can, but we don't.

We know we can. So there's two reasons. One is folks are still worried about risk, even if the AI is good, if it missteps, like in a small percentage of the time, the legal team is worried about risk. That's the number one. Thing number 2 is a lot of recruiting teams, and this is probably top-down from like the C-level, they're not caring about candidate experience the way they used to. And honestly, it's like, it's kind of short-sighted. I, but I do think a lot of recruiters are just like worried about filling recs and they're worried that if they don't hit their goals, they're gonna get fired. And everybody still has a bunch of PTSD from all the layoffs. Yeah. And, and, and so like, you know, recruiter, it's created the shift of like recruiting as an industry reverting back to transactional hiring instead of like caring about candidate experience and the long-term relationships. That's insightful. But you can't blame them. You can't blame leaders.

I'm just saying. We have all these— we had these tools before and we didn't use them, but it was because it was hard. We have the tools now, they're better and they're easier, and we're still not using them. It's like, I know, everything changed but nothing changed. I know.

So here's a cool thing though. I was talking to a TA leader at Transform, and she was sharing with me— large enterprise company, they hire tens of thousands of like, um, frontline workers, like a quarter or something like that. And, and they had leveraged AI to, um, actually automate a bunch of the process. And one of the key things that they were using AI to automate was getting every single candidate a piece of feedback about what— why, like, the qualifications, why they didn't meet the basic qualifications. It was so cool.

I've had a guest on the show before I'm going to see if I also do the recording. He's an expert in frontline. I think the two of you should connect because I think it's a blind spot for a lot of leaders, frontline workers. And he's got, he's got a really tight formula that can apply. It's not the same for everyone, cookie cutter, but it can apply to a lot of scenarios. And that's smart. Immediate response is key. So there's two things in what you just said that make it very smart. And sometimes they get missed. One of them is that's got to be right away, like immediate. Yes, yes. We talked about, you know, 3 months later, it doesn't matter what the feedback is. Shit, man. Exactly. 3 months later, why didn't you get the job?

I don't give a fuck. I know, I've already, I've already had such a bad experience, right?

Yeah. So immediate is one of it. And then the other one is, it's something valuable, constructive, productive. In this case, something else. Sorry, I just blew into my mouth. It could be something else of value. I mean, you could even do something like if you're a retail products company, you can say, hey, I'm so sorry, we didn't, you know, you didn't fit the bill, but here's a cup of coffee. Obviously, if you make coffee, whatever, like, yeah, something back that says we are paying attention. We're, you know, yes, you're number 9,999, but we value all 9,999, you know, with something like it's just, it's dehumanizing to just get a thank you for your application, never hear back again.

No, totally. And I actually think this is where your approach can come in as well. So like, A, they give feedback when somebody's not— doesn't meet the minimum requirements for the role, and they use AI for that, which is so cool. They get back to people really quickly. But then for the folks where it's like, oh, they did meet the requirements, but we filled all the roles that we have right now, They get— they send them a separate email and then those folks get put into a nurture database because they're going to be opening up 10,000 more of these roles next quarter and they'd love to hire this person or interview them at least to see if they're a good fit. And I just thought—

Steve, so you're saying it's not goodbye forever. It's no for now, but maybe later. Oh, or maybe goodbye forever. Maybe like, hey, this is just never going to fit. That's great. I want to know that. But it could also be, is it going to fit right now? Or there's a better fit, you know, like any one of those is better than nothing. But I think that there's a good little something in there to provide in that feedback. Is this never? Don't bother applying? Or is this we'd like to see you apply and be honest about it?

Yeah, no, I agree. And I think that's like something that could be unlocked for the industry in the next few years. The tech is already getting there. It's already, it's already quite good. And for some use cases, it's really good. And it was inspirational to hear the story about this enterprise company that was doing this. And guess why it happened? It happened because— no, top-down, like, mandate. We care about candidate experience. Okay. Yeah. And we, we want to give every single candidate actionable feedback. And so it's got to come from the top. So anyways, it does.

Yeah, you're absolutely right, dude. Um, an hour has gone by, and, uh, first of all, I wish we had 2 more hours. Like, I could— like, you know, I'm gonna have to have you back if that's okay. But I do have a question that everybody needs to answer who's a guest of the show. Normally I have two. I'm gonna let you off with just one though. That question is, who should be our next guest?

Oh, all right. You should invite Luke Eaton. Okay. Onto the show. Do you know Luke? Um, of? Super sharp guy. Uh, so Luke is really cool because like, all right, there's a lot of like influencers on LinkedIn. Um, many of them are optimizing a little bit too much for impressions and like followers and stuff like that. And actually, you know where that reverts to in our industry? Everybody starts making candidate-facing content and they start to talk about layoffs and like remote-friendly companies, because that's what candidates want to hear. Because guess what? There's thousands more candidates than there are recruiters. Luke started this journey, um, a few years ago where he said, I'm gonna be a content creator and I'm just gonna like not care about any of that stuff. And I'm just gonna make helpful content for recruiters, recruiting leaders, and RecOps, even though like it's not gonna get me as many likes, it's not gonna get me as many followers. And he puts out such amazing, like actionable content based on experience being in the trenches. And so he's like the most, Unknown best content creator. Have him on. You're going to love him.

So is this Luke Eaton from the UK or— Yep. Yeah. Okay. So he's at MATTER, I think, right now.

Yeah, exactly. I just had him on a little, like, you know, real talk panel that I hosted. And he's— yeah, he's awesome.

I'm definitely going to reach out to him. For sure. I'm gonna have to invite you back because I had a couple questions that we never got to. Let's do it. All right, man. Well, thanks a lot for being on the show today. Very, very special, folks. Uh, we are coming back next week, and our topic for next week is going to be, um, actually, we're going to have Rusty Roof, who is— some of you might remember him from way, way back. He was an industry thought leader back in the day who's now turned into a CEO and angel investor and advisor. Look him up, Rusty Ruff. That's R-U-E-F-F, not R-O-O-F-F, R-U-E-F-F. So it's going to be a really fun one, big shot for sure. Steve, we're definitely going to have you back for sure. And until next time, everybody, take care of yourselves, be kind to yourselves, and