EP 144: Leveraging AI for hiring with Steve Bartel, Founder & CEO of Gem

In this episode, host James Mackey, CEO of SecureVision RPO, and Steve Bartel, Founder & CEO of Gem, dive into AI-driven talent acquisition, exploring its potential for recruitment processes. 

They discuss the intricacies and challenges of incorporating AI in talent acquisition, from leveraging AI to streamline sourcing and candidate evaluation to maintaining personalized touches in tech recruitment. The conversation explores the importance of quality data sources, calibrating AI systems, and mitigating bias in the hiring process. 

   0:00 Leveraging AI for outbound sourcing in the tech industry

11:24 Calibrating AI for recruitment success

25:12 Building talent pipelines with AI

Thank you to our sponsor, SecureVision, for making this show possible!

Our host James Mackey

Follow us:
https://www.linkedin.com/company/82436841/

#1 Rated Embedded Recruitment Firm on G2!
https://www.g2.com/products/securevision/reviews

Thanks for listening!

Transcript

Speaker 1: 1:12

Hello, welcome to the Breakthrough Hiring Show. I'm your host, James Mackey. Today we have Steve Bartell back on the show. Steve, thanks for joining us.

Speaker 2: 1:19

Great to be here. Thanks for having me.

Speaker 1: 1:21

Yeah, thanks for being here again. So I would like to start off with AI. There's a lot of cool stuff that's going on with generative AI right now, and I'm really curious to, as somebody who's one of the top leaders in the recruitment space, particularly within the tech industry, I'm curious to see how Jim and just you are thinking about how talent acquisition teams are going to be leveraging AI. You mentioned that people seem to be getting really excited about using AI to reach out, to do outbound sourcing essentially, so I'd be curious to learn more about that. I think all of us are, because that's probably a use case that has been talked about a lot less than, for instance, like screening about applicants or stuff like that.

Speaker 2: 2:05

Yeah, it was really interesting. I actually ran a survey. I surveyed my network just last week about this and I asked them what do you think are going to be the top use cases that drive real value like real value for recruiting teams and maybe that was the key word, like actual value as to why people gravitated towards this? I expected inbound applicants, screening inbound applicants and ranking inbound applicants to be at the top of the list, but it was actually number three. So number one in terms of where folks thought AI would drive actual value was for sourcing of external candidates. So that's both like finding folks but also helping to draft the messages that reach out. And going back to AI use cases and what's often talked about, I do think the content creation piece is pretty widely accepted as a place where AIs can help. So of course, the drafting makes sense. But actually the external finding of candidates was pretty interesting to me From what it's worth. The use case number two and maybe we can talk about that in a little bit was candidate rediscovery.

Speaker 1: 3:12

Yeah, candidate rediscovery, I definitely want to jump into that as well, but from a sourcing standpoint, I find that really interesting. So for at least for the tech industry, right, like a lot of the outbound sourcing I would assume, are your customers referring to some kind of integration or some kind of extension that's sitting on top of LinkedIn recruiter? Is that, when we say outbound sourcing, is that what we mean or how are we thinking about that?

Speaker 2: 3:34

No, actually, I think what folks want you talk to any recruiter or sourcer in an ideal world what they want is a steady stream of qualified candidates that hit their inbox every week without the need to go into any tool and find them.

Speaker 1: 4:08

And I think what folks are imagining is something that could aggregate all of the sort of publicly available data and profiles and find the folks based on the job that they're sourcing for. The way to evaluate most candidates would be through finding their profiles on LinkedIn. I guess you could scrape different job postings, but how would they have access to those applicants unless those people came inbound to their specific company? So it seemed like if it would be leveraged for outbound sourcing at least, particularly for the tech industry, there would have to be some kind of access to LinkedIn. I'm just trying to think like what other source AI would be targeting to find a relevant person to do outreach to.

Speaker 2: 4:42

Yeah, that's a really good question. I think that there's a lot of sources of data out there and a lot of like publicly available data, even LinkedIn data. A lot of LinkedIn data is publicly available and there's a whole industry around data enrichment and like data aggregators that all they do and all they get really good at is finding publicly available data whether that's LinkedIn data or other data and aggregating it into one big like database that then vendors can plug into. There's dozens and dozens of these data enrichment providers. By the way, these exist not just in the world of recruiting. They exist in sales and marketing and advertising. There's massive industries built on top of data enrichment across all sectors.

Speaker 1: 5:33

Yeah, that's going to be really interesting. I guess the primary metric there would be qualified pipe or the sourcing quality, right, so some kind of metric to look at. Okay, I've candidate sourced. How many of those are converting into phone screens or making it past initial recruiter phone screens to primarily a hiring manager interview you? But you're right, One of our customers now is a company called Just J-U-S-T-T and they basically do something. It's like a similar data aggregation. It's a totally different industry, but basically they're leveraging data, AI and automation and they're in the fintech space. So specifically, they do chargebacks. They help companies with chargebacks. So, let's say, a large online vendor like Best Buy, for instance, if somebody has something delivered and they request their money back for the product's broken or it's missing a piece or whatever else like that, they can request the money back and have it automatically. If it's under a certain dollar amount, the company will just refund the money immediately, and so there's a lot of fraud related to this.

Speaker 1: 6:31

So, companies like big vendors, merchants, are losing millions of dollars a year, and so basically, this company just is leveraging like a ton of different data providers to put together a persona and profile around the person requesting the chargeback, to see if they've requested a ton of chargebacks, to identify fraud and then basically customize a response to the bank and to the consumer. So it's this data aggregation to essentially help companies be a lot more effective and efficient, right? So I do think we're seeing like a push from like a data and AI perspective, probably across like several industries. So I think you're right here. I'm curious to learn more about the data sources and then to like the parameters, right.

Speaker 1: 7:18

Like the one thing about outbound sourcing is, if we're going outbound, then we'd want to make sure that we're going outbound to people we actually want to speak to. I don't know, tenure would be an obvious example of some companies want two plus years of experience at XYZ companies and we want B2B, complex SaaS experience, a growth stage company with 200 employees and all that kind of stuff. We don't want to fill the recruiter pipeline with 20 phone screens of people that are not going to be quality candidates. So it's again like the quality of those data sources and then how AI actually filters through that appropriately to ensure the hit rate is actually good. That's going to be, of course, I think, the biggest challenge. If a company can solve for that, there's obviously a huge value add.

Speaker 2: 8:05

Seems very challenging.

Speaker 1: 8:07

Very hard product to build.

Speaker 2: 8:09

Agreed, that's the dream, but I think there's some good crawl, walk, run approaches here where you can get started, and then there's going to be V2s and V3s that are more fully automated. So here's just like a few examples. So first off I think you hit it on the head you need to put recruiters in full control over a lot of the criteria that the AI is using to source. The cool things about LLMs, like large language models, is the way you define those criteria can be really flexible. You don't have to fill out a rigid search interface. You can actually write text prompts.

Speaker 2: 8:45

Now, I think an open-ended text prompt where the recruiter or sourcer can write anything, that's probably not the right level of freedom and control. You want a product that actually puts some guardrails around that. So maybe there's a prompt around location, there's a prompt around tenure, and some of those things could actually be filters and controls, in the same way that you might search LinkedIn or you could alternatively have like a text prompt. Here's a cool example If you want your location to be a little bit more abstract so that you don't have to enter every single city, but you can say close to this geo, that's like a more flexible way of doing a location search, now using like LLMs, compared to previously, where you had to figure out exactly how to put those entities into the search, but that's a little bit of an aside.

Speaker 2: 9:34

So if you have a way of collecting the requirements for a role across all these different dimensions and allowing the recruiter to fully control what prompts get put in there, I think that's going to really help. So then the AI is not going rogue and sourcing a bunch of people that aren't a good fit. But you also don't want this blank slate problem. Ideally, the AI should give you good defaults for all those prompts based on your job description, so that the recruiter is just jumping in there and saying oh no, let's tweak these three, but for the most part, these other seven are right Now. The challenge then becomes until it's calibrated, you can still reach out to a bunch of people that aren't the right fit.

Speaker 2: 10:21

So I think it's really, and then you're talking to a bunch of people You're like, oh shoot, you'd never hire this person. So I think there's two things to address that. First off, I think it's really important to do a calibration phase, the first 20 or 50 profiles that the AI finds you. Let's use that to actually get better and refine that specific search, in the same way that a recruiter or sourcer would do a calibration phase with a hiring manager.

Speaker 1: 10:52

Yeah, that's what we do. So we do contract embedded RPO recruiting for tech and we do 48 hours after a kickoff call we do a calibration meeting and so for each role that's assigned to us, we bring 10 profiles to the table and share screen with the hiring manager and say what do you like, what do you not like? Would you want to interview this person just based off their LinkedIn profile? And this is before we actually send the messages. We don't send any messages until after we do the calibration.

Speaker 1: 11:14

So I guess that would be one way to do it. It's just have calibration but don't have the messages sent. Make sure that the relevancy is there. And then also they could put like blockers in place. Like ai can only schedule three interviews a week or five interviews a week, so there's guardrails, so you don't get your calendar filled up with 20 people that aren't relevant. So there could be like steps. Calibration up front didn't limit, I suppose, the number of slots that ai can fill on a weekly basis and then just slowly expand, which is like onboarding a sourcer.

Speaker 1: 11:45

It's like the same process.

Speaker 2: 11:47

Yeah, exactly. Take exactly the same process that we use to onboard a sourcer and then build those into the. You just described the V1 of how the gem AI is going to work. It's remarkable, james. Yeah, if you have a crystal ball into our product or not.

Speaker 1: 12:04

I've been in tech for too long man. It's a decade. We've helped over 200 customers hire. I've just been doing this a minute. One question I have for you, though. It could also be leveraged like the person a human being is looking at all the profiles, but opposed to doing any kind of like the outreach messaging, dropping it into AI, and I guess that would be more of like a content, like creation type of deal. But I'm just wondering, have you given thought to that too? Where is there a way I could be? I don't know, that's just might be like more content creation. I guess if you take resumes or exporting LinkedIn profiles and then totally get, you get a few hundred profiles and then you just drop those into some AI tool where it's leveraging, for instance, like with Gem, like email addresses and all that kind of stuff, I guess that would be running large email campaigns with just like importing a list of candidates or something like that and just letting it run.

Speaker 2: 13:01

Yeah, totally, and actually we already support a lot of that in the platform today, where, if you find the people you care about, you can leverage Jamai to help with the draft creation so you can have much more personalized drafts and then those email campaigns with automated follow-ups that go out. But I think one thing that's going to make a lot more easy to adopt is that, out of the gate, we're actually just going to have AI serve up the profiles and then we're still going to have the sourcer and recruiter review the ones that it finds and then that way like you're not reaching out to somebody that it wouldn't make sense to hire going to get better and better because we're going to feed those positive profiles back in and suggest new search criteria for you so that their search gets more and more calibrated, and then you might get to a point with a search where you're just like, hey, the last 20 people I've reviewed all look great. Let me click this button to put it on autopilot.

Speaker 1: 14:00

Yeah, yeah. So also, though, do you start with a specific type of role? Do you start with high volume, entry-level roles, or do you start saying, okay, we're only going to do this for engineering hires or SDRs? I assume I probably give this a lot more thought than me, but there's probably a use case where you feel like it's going to be a little easier to get started with this tech, prove out the value, and then there's more sophistication that's required for certain roles that might require a lot more specialization. Have you thought about that? What do you think? For whatever reason, high volume entry level seems a little easier to me in a sense, but I'm wondering.

Speaker 1: 14:40

Within tech, that's where it starts to get tricky, particularly in this market where hiring managers like want a very hyper specific, like now more than ever. We're getting hiring managers that are like we want somebody that comes from this industry, that's built, sold this specific solution. And because there's a lot of candidates on the market, it seems that if some executives they realize, okay, we can get somebody that's like hyper specific and dialed in, and and that's where I feel like that, like it would be the hardest to leverage ai for stuff like that so, interestingly, ai is actually pretty good at getting the hyper specific stuff right because it can go through a super high volume of candidates and filter for those very specific criteria.

Speaker 2: 15:20

I think what those hiring managers might find is that limits their talent pool a lot more than they thought. So I think it's going to be important for the AI to be able to say hey look, you've exhausted the talent pool of 30 people that fit that hyper-specific criteria.

Speaker 1: 15:35

What do you mean, we're in New York City, oh gosh, one of the biggest things I'm seeing now too, it's like they're getting hyper-specific. You're right, they don't understand the size of the candidate pool. But the other aspect, too, is even when they are getting candidates, they're burning through them.

Speaker 1: 15:51

They don't understand the pace is front loaded. You might get 10 profiles or 10 people in your process in the first couple of weeks because people are open to jobs. Right now they're more open to taking interviews. I think right now we're having a lot of luck building pipelines quickly for customers within tech, but sometimes the searches are so hyper specific that they get 10 people and they're like oh, wow, this is great. Two, three weeks like we're really building a strong pipeline, and try to keep reminding them like, yeah, ok, you have great pipeline now but you're going to burn through this, this stuff.

Speaker 2: 16:23

So you got to be careful and just realize this pace is not sustainable totally, and actually that's where I think is probably the the biggest limitation or risk with leveraging ai is for these more specialized roles, I think, because ai can feel so automated and like it's very little work, there can probably be this temptation to just mass email a bunch of people, but if you're going after a very limited population, you actually might want to take a higher touch approach where you actually review each of those people and think about and maybe improve upon the messaging that's teed to go out. Yeah, your chances of connecting with them, especially if the total population for the folks that would be qualified for a role, especially if you're being very specific, is the 100 people out there right, because otherwise you're gonna have to go back to the drawing table and remove some of those requirements, which inevitably happens.

Speaker 1: 17:14

We have so many searches right now for customers that are just hyper specific on our weekly check in calls. It's okay. Here's the size of the candidate pool. This is how many people we've reached out to. Here's all the follow-ups we've done. Here's the channels we've reached out on. Here's the ways we can expand the search. This is our recommendation and it's just man, it's constant and so you're right. Like for us, if I have 30 people in a market for a hybrid role with a super specific skillset, it's just a manual process, right? We're just literally going through every profile and I don't think we would leverage AI for those types of roles. Honestly.

Speaker 1: 17:51

I think we probably just do that ourselves.

Speaker 2: 17:53

Well, imagine if AI could help you find those 30, 50 people to save you the time of pouring through 10,000 people to find those hyper-specific profiles. And then you can take the time to do what you need to do best of like actually figuring out how to personalize and like connect with those folks. And then potentially I could be like hey, look, if you drop this one constraint from your search criteria, that would open you up to from 50 people, that would open up your total pool to 500., and they can know that because it actually has canvassed all those people. And then that helps you show up as a strategic partner to your hiring manager and say, hey, are you open to dropping this one constraint? It seems like a nice to have. If we do that, we can 10X that the available talent.

Speaker 1: 18:41

Right, yeah, yeah. Okay, I think I had an initial question. Then I rambled on with a bunch of other stuff. Is there like a specific role or function where?

Speaker 2: 18:50

Yes, yeah, good question. Yes, I think you're right. You hit it on the head. I think it's, for starters, like AI is going to be most effective at, like evergreen roles where there's a pretty big talent pool for them, roles where you're going to be hiring a lot of those folks. So good examples are like software engineers entry level.

Speaker 1: 19:09

I feel like you're ready to go there.

Speaker 2: 19:11

Yeah, totally Entry level roles, SDRs, account executives unless it's like enterprise account executives with very specific credit deal.

Speaker 1: 19:22

I got so much of that going on right now. It's just driving me crazy I know.

Speaker 2: 19:26

But like on the flip side, like a head of trust and safety with very specific criteria, that might not be the right role to throw like a big ai sourcing or like executive search right too much automation around that stuff totally, and maybe ai could still help canvas the market and help you put together that docket of 30 50 people.

Speaker 1: 19:48

That, yeah, exactly like how much available talent do we have that like kind of talent insights perspective where it's helping in the planning phase? I think is another huge value proposition, right, like how can we pull in a bunch of data in the prep process? It's like the whole understanding of most of the work occurs before the search opens. I remember, particularly during COVID, we lost so much revenue At one point I was billing myself out as a headed delivery for an applicant tracking system called CoMeet.

Speaker 1: 20:20

It's still out there. I don't know if you've heard of CoMeet, but they work with a lot of SMBs, early stage companies, and they wanted to roll out like a fractional embedded recruiting model alongside their applicant tracking system where, like, basically customers could click a button and then they could engage this on-demand service. So I was running a delivery for that company. I took on another CPO role for another company and then I was doing like fractional head of TA for another company. It was nuts. Basically, one of the first things I would implement, like when I would go into a new company, is, like what is your pre-job opening process? So what steps do you have to take prior to a role being open? And we ended up developing like a nine-step process. It was like pretty robust in order for a hiring manager to actually open the role. There was a lot of things that we required to go into planning, because I always said that most of the work actually happens up front and if we structure things in the right way, we'll have a lot of momentum going into the search. And then we also had guardrails in place, even once the search was open, where, if criteria changed so much, we'd actually require the role to be closed and we had to restart that initial onboard process ahead of time. And so I think it's like we look at every role as an investment and to make an investment, we have to have data and we need to understand really what we're getting ourselves into.

Speaker 1: 21:39

And yeah, I think that there could be definitely a large use case for talent insights prior to roles being opened and having that part of like potentially even approval processes before roles are open, to really understand, okay, what's the investment we're making, how much time, resources, and then also that can also help plan like capacity, like where are the bottlenecks going to occur? Are we going to need coordinators, because it's high volume and we're going to have a ton of candidates in process. Do we need sourcers? Because it's a harder role to source for and we're not going to necessarily have enough people in process. Do we need, like, more full lifecycle recruiters, because these are like more senior strategic roles? And how do we think about segmenting that team? How many hours of interviewing are going to go into that? How much time is going to be required from the hiring managers? Do the hiring managers have enough time to support the hiring plan?

Speaker 1: 22:23

There's a lot of talent insights when it comes to, I think, capacity planning and then also the size of the talent pool.

Speaker 1: 22:30

All this stuff can be done upfront and, quite honestly, it is quite a manual process that requires years and years of experience in tech.

Speaker 1: 22:37

I'm just thinking about the application from a talent executive perspective, because I spend a lot of time in spreadsheets putting together capacity plans and proposals for customers on the amount of headcount they're going to need for the town acquisition team in order to deliver on their hiring plan. Yep, and it goes into like everything about hours required, different roles, responsibilities, where the bottlenecks are going to likely occur, based on the types of roles that they're hiring for. It's like really robust stuff, but it's a process that takes me like hours to put together. Like I have to request a ton of information and it's just very time intensive, and so I think that there would be a very beneficial use case there that should be leveraged even prior to not even just to correct course once a role is open, to say, hey, here's how we can change the criteria, to have a lot of those insights upfront, which would be my dream as a talent acquisition executive, quite honestly.

Speaker 2: 23:28

Totally, and I think this is where AI combined with a very strategic search firm or RPO or staffing agency, or internal teams right.

Speaker 2: 23:38

Or internal. You're right, yeah, totally. But we need to take the recruiting expertise of a firm like yours and embed that and hopefully help automate a lot of the busy work of collecting the data and doing the research, and then have you make sense of it, that talent pool, and then show up as a trusted partner to the hiring manager use case, and we also wanted to touch on leveraging ai to find candidates that are already in the database, which I don't know if this is like a similar process, now it's like easier because you already have all this data in the system.

Speaker 1: 24:11

Is there anything you want to dive deeper in on that topic?

Speaker 2: 24:13

yeah, I thought it was just interesting.

Speaker 2: 24:14

I thought it was interesting that these two use cases showed up above inbound ranking and matching just because, like for the last five years, I think, a lot of people have been talking about inbound ranking and matching and so to see these two use cases get almost like two to three X the number of like votes in this poll that I ran was just really surprising to me, and I think that it shows potentially like the power of like generative AI and like some of the new things that it's unlocking compared to what was possible before.

Speaker 2: 24:47

But yeah, no, I think people are really excited about that because, especially for larger enterprise organizations that have millions of applicants in their ATS and if they've been leveraging a CRM for some time, they might actually have even more records in their CRM between all of the people they've ever sourced that they've identified could be a good fit for their company people who have attended events, people who have subscribed to talent communities, right, all of these different places of collecting these talent relationships. So the idea that when we open up a search, ai could first take a look at all the people that have engaged with our company, our brand, all of the hand raisers, plus all the people that their team has identified, could be a good fit through sourcing historically versus going out and sourcing cold. Intuitively, I think that just really resonates with a lot of the market.

Speaker 1: 25:42

Makes a lot of sense to me, and so one other application I wanted to talk about, the AI note takers, where basically there's tools popping up where it's essentially putting together kind of high level data points from an interview and then basically structuring it so you have better notes. One of the issues that I think every talent acquisition executive and recruiter has is getting hiring managers to fill out evaluation forms after an interview. Right, it's just a total pain, right? Because hiring managers like a lot of the times their capacity, when being measured by executive teams, doesn't take into consideration the amount of time it takes them to hire. So it's like they have a full plate just based on the core function that they're running, but then sometimes upward of 50% of their time needs to be used for talent acquisition and there's just not enough time in the day.

Speaker 1: 26:30

So we try to do things like say, hey, just jump off the interview five minutes before the end of the time, block, fill out your evaluation, because then it's been two days, they forget 80% of what they talked about, and then all you have is do you want to move the person forward or not, and that's basically the level of insight. So I wonder, from a evaluation perspective, if it not only can take notes, but if we can get to the point where the AI kind of ranks the candidate and helps accelerate the evaluation, where maybe it's even just putting together a draft where a hiring manager can just like, review and click yes, versus filling out all the evaluation criteria. That might also be an interesting use case, what we're looking for in the role and then we can move people forward in the process without getting a form filled out in GEM or Greenhouse or whatever applicant tracking system a company's using.

Speaker 2: 27:29

Totally, and I think this use case sounds really promising to me. It's one that we're paying close attention to as we bring our ATS that we just brought to market a few quarters ago up market Because I think, especially for larger companies where you have pretty extended interview loops, your hiring manager is super busy. That resonates with me. When I was at Dropbox, I was spending 50 plus percent of my time at any given time on hiring. Who knew being an edge manager was going to be actually half 50 percent recruiter? Half your job is based on something you have no formal training with.

Speaker 1: 28:01

That's the half 50% recruiter, right, half your job is based on something you have no formal training with. That's also the part that's hilarious to me. It's like we're hiring these leaders and it's okay, what's their engineering background or what's their revenue background, and they have the experience XYZ, but then it's like half their job's hiring and we're not betting for that when we're actually hiring these executives. Right, it's just kind of funny. It's like a big miss. It's a huge miss on a lot of the times right.

Speaker 2: 28:28

That really resonates with me. But I think your point around summarizing the interview notes even that takes a lot of time writing up notes and you can't really be present. So I think interviewers, if they're jotting down as notes in real time, maybe they're not showing up in the way they would want to. So I think even speech to text, that has gotten like way better in the last year or two through generative AI. And the cool thing is there's like a precedent for this in the world of sales with companies like Gong in terms of like reporting sales calls and man. We use Gong internally. When I go into Gong and I see the magic that exists in terms of being able to nearly perfectly transcribe calls and then summarize them into an eight point outline with the key takeaways, action items, I'm like holy cow, yeah, of course this is going to revolutionize recruiting. The exact same tech would be helpful.

Speaker 2: 29:23

I think the thing we need to be really mindful of in recruiting is anything that starts to evaluate and assess candidates we want to be careful of, because AI can bring bias into the process.

Speaker 2: 29:39

After all, it's trained across a lot of human data and humans have bias right. So I think the key with anything related to AI in hiring decision-making is to be really thoughtful about how you build that technology and actually legally you have to get that audited in a number of states like New York, for example. So any vendor that's building that into the hiring process would actually have to get that audited. Now, if you're just taking notes and summarizing the call, I think that's totally fine. But as soon as you're recommending an assessment, move forward or don't. That's when you want to be extra careful. But honestly, if it was just taking notes and summarizing a bunch of the key points that came up, that'd be huge and also that might help with the problem of hiring managers taking two, three days to fill out their feedback. First of all, they might be a lot faster if a lot of the notes are taken for them and summarized for them.

Speaker 1: 30:35

I would hope so, but I still feel like perfect notes summarizing everything. They're still going to take five days to fill out.

Speaker 2: 30:42

To hit click perfect notes summarizing everything. They're still going to take five days to fill it, to hit click. But here's the brilliant thing is, even if it takes them a few days, sure, it's not going to be as high quality in terms of them being able to write their assessment, but they're going to have all these great notes they can refer back to.

Speaker 1: 30:55

Yeah, so they know like which candidate is which. We had a customer man where we had one of the searches for the same role within the company. We had three people named Michael in the search. It's like refer to everybody by last name. There's just stuff like that where it's like having the notes are so important because you don't want hiring managers to confuse people, right?

Speaker 2: 31:30

Totally, and then also just to forget a lot about what was talked about. And having those notes is really good for like recalling oh, that's what was said, that's what was talked about. Now, I remember this is how I felt about whether they'd be a good fit.

Speaker 1: 31:43

We're coming up on time today. Steve, I just want to say thank you so much for coming on the show and sharing your insights with everyone. It's incredibly valuable.

Speaker 2: 31:49

This has been awesome, man. I love the conversation around AI. I think it's going to be game-changing for the industry and, yeah, looking forward to talking more soon.

Speaker 1: 31:57

Yeah, absolutely Me as well, and thank you for everybody tuning in.