Who Owns Your AI Agents? [Ft Laura Stapleton, Foley]

We want to do this AI transformation, there's this added layer of complexity to reimagine how work gets done. Have you come up to where you felt like you needed to talk to them about AI is not coming for your job? It's not about cutting people. Like, it's how do we want to organize ourselves as a result of the changes in the way that we work. The idea that I think companies should be thinking about is like, you have those four ICs who are A players, they should be managing agents. There's a huge layer of ownership and accountability that the agents are never going to take on. How are you going to have enough domain expertise to have the stewardship so that you can drive accountability at the human layer? Welcome to AI 2030.

We have Laura Stapleton with us, Chief People Officer at Foley. Welcome. Great to be here. Thanks for having me. We've known each other for a while, so this one should be fun. Yeah, I'm really excited. You are one of the, not one of, you are the most unique Chief People Officers I've ever worked with, mainly because the way you look at things is just so different from a typical person who kind of came up in, like, the HR side. Why do you think that is? Oh, I think probably because I came up through talent. I used to say, I'm not an HR person, I'm a recruiter. I was super passionate about talent and driving performance. So I came up on that side and was sort of like, the benefits, the compliance, none of that was me.

It is now. And I care very much about those things, to be clear. I really, I think coming up in staffing, you know, where it was super cutthroat environments, I was really, like, I would say, I had more of a sales mentality than anything and really grew a passion for performance and talent on top of that. And that's really what, when I say, like, the most unique, I feel like there's so many HR leaders out there that are, I don't want to say scared of letting people go or ensuring there's high talent density, but there are not as many that I've found like you that are so rigorous about ensuring that the talent level is high enough to be there that I I feel...

You're not you're not a... How would you call it? You're not protective of the people that that you're like, well, I hired them. So it's like, if they're not performing, like, we have to to level them up. And if we can't level them up and we're giving them direct feedback, we have to move on. Yeah. I mean, it's not fun to underperform. So I don't think... I mean, I would say I'm a business first people leader, so I'm always acting what is in the best interest of the business, and that impacts all of the employees. So you have to think about it collectively. And I feel like when it comes down to the individual, like it's miserable to be underperforming. Nobody enjoys that. So if it's genuinely not the right fit, I would rather make that decision and position that person to go find what lights them up where they're going to be happier and, you know, as a result, more successful.

Yeah. So... In in the role I'm I'm in now in Cadre, so interesting because, like, while I feel like I can add value on the operational side or on the brand and demand side, one of the biggest value adds that I think I can I can provide to the business is the people side? It's like not being the is being the person who's gonna help level them up, show them how to do something, watch them do it a couple times, and then let them go do it and trying to keep level them up. Do you feel like like that's where one of the biggest, like, value adds you're able to provide to a business is making sure... I almost think of... We've always joked when we worked together that you're like the Wendy from Billions. Yeah. That just, like, the coach is gonna come in and, like, get these... Like, understand what these people really care about and figure out how to level them up and like get them them on the right, you know, track?

Yeah. I think it's also giving people the space to do it. Like I'm a big believer in ownership and accountability. So you give people opportunities and like let them run until they can't or they get blocked. And then I really view my job as unblocking people. If you have the right person with the right level of self motivation, I think people are usually capable of far more than they think that they are. So it's just putting them in the spot to stretch. And that can be really uncomfortable, right? So that's the hard part is like, I deal with a lot of discomfort on my teams as a result of that and constantly pushing people or putting them in spots where they're like, woah, I need you. So I'm also super careful and intentional with like when I insert myself. And when I'm like, nope, you're fine.

Just like push a little bit more and you're gonna get what you're trying... Oh, you know, people have to work through things on their own. So I'm a big believer in that. Yeah. And you've been at Foley for a year now Yeah. Or so. Right? What's it... Well, first off, tell me what does Foley do? Yeah. So Foley is an end to end platform. So we help organizations recruit onboard and continuously monitor safety sensitive employees, so mainly drivers. So really making like compliance, it's super costly and oftentimes a headache, we really turn that into more of a strategic advantage for our clients. So I'm a company, I have a fleet of vehicles. I need drivers. I need the drivers to be compliant. I can't... Mean, I think we were talking earlier, you said it was like up to like $10,000 a day or something when a It fiber falls out of can be like insanely such a cost drain on organizations if they fall out of compliance.

So we handle everything from like background checks, drug screens, motor vehicle records, like the list goes on and on dealing with like the DOT, so on and so forth. So it's a very heavily, compliant driven part of the the universe that we're supporting. And how have you seen AI creep into the people space while there? Because I would imagine in the last year, things have been going... I mean, AI is just advancing like like crazy. So how have you guys been seeing it progress? I mean, it did not creep in. I, this was the first time that I joined a company as a people leader where AI was on the table. So I've joined as the head of people for organizations before, as the first head of people for organizations before I've done that. But this was the first time where it's like, I can't come in with my typical playbook and be like, boom, problem solved.

Like that was not going to happen because they're saying like, we want to do this AI transformation. So there's this added layer of complexity to reimagine how work gets done. And that goes far beyond just like, or design, or it's not about cutting people. Like it's reimagining the way in which we work. And so that sort of mandate, if you will, was made to me like throughout the interview process. They were like, here's what we want to do. And I had to openly say like, I've not done that, right? But I'm super excited and I'm leaning in and I'm really passionate about like the way that work is changing and I want to be sort of front and center as the world sort of shifts in that direction. So it really, it was a very unique experience for me in terms of like, how do I come in here and help this company with this transformation?

Well, also just like getting foundational HR up and running. It was kind of like two different tasks at once to be like, do we have sort of a modern enough structure that we can then like layer AI on top of it? And that's really what we've been working through over the last twelve months. So I think a lot of people leaders and HR leaders these days are thinking about AI in maybe more basic ways. They're like, how can I use AI to screen candidates? How can I use AI to, you know, give me scorecards based on, you know, the interview process, based on what we wanna look for in that? How have you guys seen it? Like, how are you using it? Are you using it in those ways or is it going beyond that? I mean, way beyond that. That to me is like table stakes now, if I'm being honest.

And a lot of it, it's like you get into the build versus buy conversations. There are great tools out there depending on like what ATS you're using. Like a lot of that at this point from my perspective has been solved. So it's just like figuring, deciding that you want to do that and then doing the research and going and doing it. So like that is all happening on like the hiring front. And talent acquisition, as I'd mentioned, like that's my bread and butter. So I would say like I came in already. That was the spot where I had a lot of opinions and I had more experience because at my previous organization where we were working together, we were starting down that journey already. So I'd had like a couple of months of like the using agents for the sourcing and, you know, the screening, the notes, the intakes, all of that as it relates to hiring.

So that part was like sort of step one, you get some quick wins, you earn a little bit of credibility on top of that, like great, we're moving in that direction. But my role has really changed in terms of like thinking about the organization as a whole. Like how do I help and how do I think about this entire structure versus just like the functional areas that I own. So I'm thinking about it like much more holistically. And I think that really started with like access. You know, that was where we were at when I joined, which was last August of like access, like what tools should we use? Who should have access? Do we want to tier that? And I don't even think we were really thinking about cost at that point in time.

It was just like, should we have GPT enterprise or should we, you know, like what do we want to have and who has access? So it was really starting with that. And then, you know, from there we sort of shifted into like, okay, now we'll look at usage. People have access. How are they using it? And now we're hyper focused on impact. Like where is that impact? Because kind of like, I don't want you using it if you're not having impact. Like something's disconnected. Like if we're not having the impact that we anticipated that we were going to have or if we were like, were we even clear with what that impact was supposed So to that's I would say the stage we're at now when we think about our AI maturity curve. We have been seeing almost the exact same thing at nearly every type of organization, SMB to even some of the enterprise that we work with, where they're like, wow, we've been investing in it.

We've given access to the claw that we picked, you know, whatever LLM we're gonna go with, and then... And now we're saying, alright. We have the harness. We have the right LLMs. People know how to use it. They're actively using it. But there's this, like... You know, they say there's, a bell curve to things. We see it opposite. It's like a barbell. It's, big on both sides. Like, the people who are barely using it to the people who are power users. And in the middle, there's almost nothing. Yeah. Because when people start using it, it's like brains explode, and they're like, I'm gonna use this for everything. Or they start using it, and they're like, I don't like this at all. I'm not doing it anymore. And so there's very few people in that middle because they just... Once they start using it, or they go back.

And and when you think about the the impact side, it's like, alright. So the... There's a lot of people using it. They have their daily jobs. They're now building skills for, like, everything that they do. Are those skills tested? Are they just rolled out to the organization? Do you have them in plug ins? Do you have, like, a marketplace where people can go, oh, based on my job, these are the 18 skills I need, and they've been vetted? Just to your point of, like, do... When you go and set out to say, like, hey. I'm gonna go build something, have you mapped out the things that are gonna be true at the end of this? Like what impact do we really wanna get out of this? Because people are spending more time now building skills for things that they may never even use again. Right. And it's like, just do the thing. Yeah.

If you didn't build the skill, you would have done the thing faster. Right, Right. We saw so much of that. And it's more like it's not even what we saw. It's what we don't see that keeps me up at night of like, you know, when you talk about cultures of high performance and, you know, transformational organizations, like what if somebody leaves and they have all these agents running and I don't know about it? That's right. Like what about that? So I think it's really important to then like take a step back. And I feel like we went sort of on this like phase of exploration and now we're getting like really way more intentional with the strategy and how do we want to organize this work and how do we want to organize ourselves as a result of the changes in the way that we work.

So that will help us ensure that the quality of what we're doing is there, that there is sort of intention and strategy and that, you know, the cost is going to be worth it. And we have really like thought through versus like go build anything, use AI because we just want you using it. Like we've now, I think, graduated beyond that to the point where like, okay, let's be super strategic and intentional and make sure that, you know, what we're doing is super high quality. How many people at the company? About two sixty. Have you done any trainings or have you come up to where you felt like you needed to talk to them about AI is not coming for your job, or maybe it is. Because we have... What what people... What we're seeing is that a lot of mid market businesses and people who work there are very nervous about AI coming for their job because they're seeing news from, like, Oracle or some huge massive company that is scaling AI agents like crazy and potentially...

And laying people off as a result of it. And then in the mid market, they're like, wow. This is gonna happen to us. We're just... Honestly, from our side, we're not seeing that. Yeah. We're not seeing, like a lot of people saying or companies saying these people are going to go now. It's more like they have a different job now. Yes, exactly. That's the jobs are changing. So I really don't think about it from a headcount perspective. Earlier on, like last year, I probably did. I probably, you know, had the thoughts, I definitely did, of like, how is this going to impact, like how many people we need? And it's really not as much about that because I think you could see growth in some areas of the business. Like if you have AI really accelerating outputs and outcomes that are driving revenue, I want more of that, right?

So I think we'll see growth in certain areas as well. But it's really about changing the jobs itself. And we think about that in terms of stewardship, right? It comes down to like the context and who is responsible for what the agents are doing. There has to be that level of accountability, and we do that through stewardship. So we're in that process now of defining what that looks like and what that means and really inserting that into jobs. And what's interesting is the people that are best suited for that have the domain expertise. It's not necessarily the director in every case. So it's not as easy as like every director is now a steward of this domain or of this sort of subdomain or whatever it may be. We're finding that it's like, who normally answers that question?

Right? And like that person is probably best suited to be a steward of that piece of context or So that piece of we're modeling all of that now. That's cool. And that I think is going be really interesting and where we'll start to see those jobs shift. And for the people that are leaning in, like, they're going to be fine. Right? It's when you're so afraid that you become paralyzed or you don't want to lean in because you're like, this is going to take my job. It's like, this is happening. So your job is probably going to be different than what you imagined. It's going to be myself included. But if you're leaning in and you're on this journey, like you're going to be just fine because you're going to build up a completely new skill set in a lot of ways that leverages the expertise that you've built, you know, to drive efficiencies through agents.

So that's the path that we've been on. And, know, we have an example like our people operations manager. She recently shifted into a knowledge and context enablement program lead, which is not, I think, where she thought her career journey was going to take her, But that's as a result of leaning in to AI and really being somebody that, again, like, we don't have it all figured out. But she's somebody that has enough of the technical chops as well as the curiosity. Like, she's super curious and as a result of has gotten herself involved in so many different projects and initiatives. And that's where like with the training and enablement, we candidly haven't done a ton of that. I'm just not a huge believer that sitting through some sort of curriculum that we built on some platform is like going to make a meaningful difference.

It's getting people involved in projects and getting them like actually doing the thing that they're going to learn. But on the flip side of that, to talk out of both sides of my mouth, I also realize that people learn differently. So I think like our future state probably has a combination of both, where you'll have content or different channels in which you can go to get more information that would make you more effective. But we do have our value creation office and that VCO hub that we've created is a place where people can go for support. So it's more like in the work support because we want to be investing that time in people that are actually like proactive in leaning in and transforming their work.

We found one thing that has been helpful is we'll do... At first, were doing, like, Claude 200 through 700 or, you know, OpenAI. And we would then realize, like, man, once you get a little bit deeper in there, people have either caught up and like they're they're even further Yeah. Than than you wanted... Or people are like still behind and they need to go over like one zero one or one zero two again. So we found that doing a couple trainings on the basics to make sure they all understand it and then doing a department hackathon is a really interesting way to go. We'll typically spend three hours, and we'll say, alright. First hour is let's all talk about the things that you do on a daily basis and what you think now that you know how to do the prompting and all that, what could be automated. You spend another hour, like, go build. Yeah. And we would just walk around and kinda help them, and then it's like, present what you built.

And by the end of it, they usually end up with a handful of skills that were like, alright. Roll this out across the team and start using them and refine them. So that's been a good good way to do kinda hands on. But the the interesting thing about what you said about context, wanna go back on, is you have somebody now whose whole job, it sounds like, or a large percentage of it, is providing the proper context for the AI to do its thing right. Is that right? I would say their job is to enable leaders to make sure that their context is accurate. Like they don't own the context in this role. It's really enabling like the structure and the building out of those domains and the ownership of those domains. So really like PM ing that effort.

That's interesting. So where we found the biggest gaps between people being like, wow, AI is transformational or AI kinda sucks, all comes down to the data and the context. Yes. Right? The models, the LLMs, they are good. They're getting better and better and better. But if you're not giving it the appropriate context from maybe unstructured data like transcripts or PDFs or any of that to the more structured data that's sitting in your ATS system If and you're not giving it the access to that and being able to build the data models to help it understand or the ontology or the Rag to help it understand what's happening in there, people are going go, this AI sucks. If the AI doesn't suck, the context sucks.

Right. Right. And that's where you're going run into huge errors. So that, I would say, is the stage that we're at now, is saying like we have to do something about how we're structuring in the context and where the data lives and centralizing that. Because once you start to have that drift and you have data living in various places, like it's going to be a mess and the agents aren't going to be able to recognize that. In terms of building out the ontology, it's really making sure that you have enough of a thin layer that's consistent across the organization so that we're all defining things in the same way, both the humans and the agents that are using pieces of information. So we're like deep diving this quarter into getting that all tightened up so that we'll be much more effective in our transformation once we have that like really nailed down.

Because we've been piecemealing it. So it's like, what are we trying to build? Let's make sure the context for this one specific thing is in place. But now we're taking a more holistic approach. The other thing that I think is interesting about the context side is that you've got... Let's say you have a team of four people, and they're A players. You got four A players, and they're great. ICs, individual contributors, not managing anybody. You gotta scale that team up. What ends up happening typically is maybe you take one or two of those leaders and you say, you know what, or these ICs, say, you're gonna become a leader. I want you to hire one or two people under you to do these jobs because you're really good at this thing. We need to scale up what you're doing. And that person's never been a leader before. Sometimes they want it. Sometimes they're good at it.

Sometimes they're not. They never hire people. They never coach people, they never fire people. Right? And so the... What ends up happening is typically the work slows down a little. The a player who is an amazing IC maybe becomes a b player because they're not doing as much and, they're coaching and they're not coaching as well as they could. Maybe you get lucky and you get one under them that's an A, then you get maybe one as a B or a C. And now things start to slow down and it gets worse. Right? The idea that I think companies should be thinking about is you have those four ICs who are A players, they should be managing agents. They can stay at that IC level, but their job changes because now they need to know how to understand...

They need to know how work with AI to build those agents and then give it the appropriate context and make sure the output is good. Right. Right? So they can stay at that level where they're cranking and crushing things and they're doing great while managing agents instead of managing actual people. I think it's an opportunity for everyone to just expand their value to the business by doing that versus looking at it as like, now AI is going to do the jobs that I was doing, I'm not valuable. It's like, no. You're even more valuable if you can use AI to multiply yourself. Right. And there's a lot of work there. Like it is a lot of work and that's the stewardship that we're talking about. I think it really accelerates people's growth. But even like in defining stewardship, you've probably had this, right?

You're working with agents and all of a sudden you have so much information in front of you, like who is responsible for ensuring that that information is good, that that context is accurate and that it's being maintained, especially in an organization like ours where like we're dealing with compliance, we're dealing with all types of rules and regulations that like we have to get right, both for ourselves and for our clients. So there's a huge layer of ownership and accountability that the agents are never going to take on. So that is where we're like really thinking about that stewardship as written right into the job descriptions. And we're actually going to be like putting out new JDs for people that have stewardship. Like if you've been assigned a steward, this is what it means. You are signing up for this. Like this is now yours. These agents are yours.

And it's super exciting, you know, and I'm really, we'll learn a lot from that experience. And I think other organizations are going through that as well. But it's exciting to like be at this stage where we're really formalizing it because I think now like you've had that experience like, my Claude did this, And it's like, no, no, no, That is just not going to fly. Driving that culture of ownership and accountability the same way that you would in a fully without agents, it still holds true. That's a big part of the transformation as well. Like ownership and accountability are now different. And there's also something a little bit scary about that for people as well. It's like, oh, I don't know if I want to be responsible for this agent.

Right? And so it's really defining what that means and how do we help people navigate through that when ultimately there will be learnings, there will be missteps, like anything that's new. So creating the space for that I think is really important as well. And then also it allows people to be more strategic. Know, because you think about something even like a board deck generator, right? So you're gonna make a board deck every quarter, whatever it is, you are going to, let's say you're ahead of people, right? You're gonna go look at your data, You're gonna be like, what's the story I gotta tell here? Right? What did we talk about last time? It's a story I have to tell. Okay. This is what I wanna go with. Alright. So now I'm gonna go build out some charts and graphs.

I'm gonna build the slides. I'm gonna think about it. I'm gonna... Okay. Yeah. And then I'm gonna send it to somebody, you know, whoever, maybe it's a CEO or president, whatever, who's gonna review it. That person has the unique advantage of being able to look at it, kind of see the forest through the trees. Right? They look at that deck and they're like, I don't think this part is right, the story to tell is right. So now you go back to the beginning, you update the data, you update the charts, you come back and all that and you pass it along. It's harder for you to catch that because you were at the beginning trying to figure out the data and figure out the story. You have an agent that understands the board deck, understands what the the structure is, understands what the outcomes of what you're trying to achieve with it, and they have access to the data and the context is good, then the agent can build the board deck and send it to you as that that person who would normally do it.

And you now have that opportunity to see the forest through the trees and say, this is not the story I think we should be telling. You go give it feedback, it does it again, iterate or iterate until it's right, and then you pass it along. So I think while the output might be the same, the job changes. It doesn't mean that you're not spending any more time on it. Just because you have an agent doing something does not mean you spend no time on it. Well, you can spend more time. You might spend more time. Right, right. Over time, it's gonna hopefully get less as you fine tune it. At the beginning, especially, like you have to give it that feedback. It's your job now to provide the context and the feedback to get it right. And that's where the context and the data becomes so important. It's scary because most things come out looking like a finished product.

So it's real, a board deck is such a good example. Could put in something and it's like visually, it's amazing. And it could be like garbage, So that is where it really comes down to like the context and the data and having that attention to detail and having that sense of ownership over such thing that like you're gonna make sure that it's actually right. And that I think is also a shift for people where you see stuff come across. I've seen stuff come across my desk. I'm like, you're kidding me with this, right? Like, or it all looks the same. And I think that's also an interesting point that people have arrived at is like a lot of the language looks the same. A lot of the decks look the same and people I think are getting even a little bit fatigued. So it's like, okay, how what do we do about that?

Right? So I think there's a lot of work to be done, and it is you know, it's shifting really quickly. But to your point, I think initially, you could spend more time. It comes back to the context and the data. You have to have ownership of that, which is why for us, like building that out right now is such a critical inflection point that we're at. And I think how we do that is going to make all the difference in terms of how fast we're able to move with our transformation. What you alluded to earlier is what you call AI slop, right? Is Yes. The Yeah. It is wow. I'm able to create this crazy looking thing. I'm gonna now give someone else the homework to go read it. Right. That's the most annoying thing I think when... And I think there's so many people that have... Are experiencing this where somebody's like, wow. Look how fast I made this thing.

Here's a six page artifact, HTML artifact that explains this SOP or this thing I wanna go do, this project. And I'm like, I'm not reading this whole thing. No. There's no way. If you can't get this onto one page... Because if you were a... As a human, you would not make it six pages. If you as a human did this. You would not. And so now you're giving me the homework to go spend extra time to review all this. And by the way, I know we're an AI company, but if I feel like AI did something that I'm reviewing, I'm like, you did not do this well enough. I want AI to do it, but it's your job to make me feel like you did it. Yes. Right? Exactly.

And so those headlines that like... I don't know if it's just... I mean, I see it all the time, it's like, two of this, none of that. Like headlines on the top of slides or anything, like 15 errors, zero in the room. Yeah. Right. Right. This is... You didn't write this. Stop it. Right. I'm not reading anymore. Like, you go back and and and make this better to the point where I don't think AI did Yeah. And now I'll go and, like, review it. I think that's also what we have. It's like that, you know, whether it's 60... Whatever you wanna call it, 60% or 90%, like, you can get to a very close end product very quickly. That final mile of refining it and making it good, that's where all the time is spent. Exactly. Exactly. And that's what I think I'm with you, it's most frustrating thing.

I'm like, you would never say that. Or even if I get something, an artifact that I'm working on, I'm like, I would never say that. Like, it's so infuriating. Yeah. Because again, like, it's that last mile to your point. So... And I think that's where it will evolve of like, how do you make that last mile Yeah. Easier and faster is like where we're all at. But a lot of people just skip that last mile. Yeah. Yeah. Which is amazing. And it's so annoying to get to to be told, hey, I got... I did this, review it. And like, could... If you had just done it as a human, yeah, would have taken me longer, but it would take me so much shorter time to review it. Right. And now you're you're forcing me to spend more time reviewing this and giving feedback that you're just gonna feed back to the AI. Yeah. You know? Yeah. Exactly. Exactly.

Well, there's no point in having all of these policies and documents if no one's gonna read them. Yeah. And people aren't gonna read them. Right. Everyone's attention span is, you know, so it's like we've gotta make it better and something that's actually gonna be applicable. Yeah. At this point, does everybody in the organization have access to one single LLM? Yeah. One one harness like Claude or OpenAI or... We had multiple. So I wouldn't say it's universal. Everybody has to use this. Like, everybody does have access to something though at this point. And depending on your usage and how you're using it, that's where we're really starting to get intentional with, like, tiers to ensure that people... I mean, there's nothing more frustrating for me to hear than someone that I know is a power user. Like, oh, I used up all my tokens and then I had to wait for like two hours to get unlocked. I'm like, what? Why?

Right? And so those are the problems that we're now solving to ensure that the right people have the right level of access based on what they're trying to do. Because I don't think it's one size fits all. Even in my own experience, right? It's so quick to be like, oh, like Claude is awesome and I abandoned GPT. Then it's like, wait, GPT is better now. I'm like, wait, you know? And so trying to manage that I think is also important for us. But I never want access to be the reason why people can't accelerate. So we're very passionate about that. Yeah. We do something similar where we have... People have limits, but we have triggers to be like, you know, someone hits it, we're gonna go and review real quick and open it up. But it's really just that telemetry view to be able to see, like, is it because they're just, like, maxing out fable for for something that they don't need it for?

Right. And they could be using Sonnet? Like, then that's a coaching moment. So you don't wanna, like, open it up and be, like, everyone gets whatever they want. You wanna stop them, but it's really more so that you can just take a peek and be like, are they using it, you know, the right way? Exactly. Are they are they just... Or are they just burning tokens for, you know, the hell of it? Yeah. Yeah. No. We're with you there. And we do have a place where people can go if they get blocked or for whatever reason or if they wanna request access for something else that we don't currently have. And we try to make sure that we're like turning those around very quickly. Or if I'm aware of somebody that's put on, do tiger teams. So we'll have a certain initiative and people are put on a tiger team. If you're on a tiger team, then like we're intentional upfront with like, what do they need? Let's make sure everybody has what they need to be able to lean in and be successful at this sprint that they're about to go Are you guys thinking much about the legal implications of AI?

Meaning Yes. Because I was talking to an AI lawyer recently, and it was a few months ago. We were on a panel together for San Diego Business Journal, I think. And he was saying how there was a guy who submitted, and I'll get some of this wrong, I'm sure, but he was you know, older gentleman, submitted 200 or something applications at like 2AM, right, on a Saturday or something like that, and got rejection letters from some of them almost immediately or on non working hours. And he sued not the company.

So he said, oh, AI is rejecting me, because clearly a human was not going in and reviewing these at that time. So, hey, AI is rejecting me. Didn't go sue the companies that did it. Sued the underlying tech, the ATS system that was doing the application reviews and won. And so there's these... I mean, it's just crazy how we are jumping all into this. There's going to be lawsuits like this that then all of a sudden go, woah. Let's pause for a second. We do want AI to help review applications, but we also want a human to go and review those as a final step as well.

Yeah. Right? And then there's some ways about like, can it review for background or does it... Can it look at... Know, can it infer age and potentially review against that? So that would be like an issue. So how are you guys thinking about the legal Yeah, it's interesting because you're using one use case. I mean, think we're thinking about legal when you first asked, like more broadly in terms of our own governance internally across the board. That we have, we can talk about that separately. But as it relates to hiring and reviewing resumes, for example, it's interesting because you've been able to auto reject based on predefined criteria for as long as I've been recruiting. So like, I'm not going to stop doing that.

Yeah. Right? And so the tools and technologies still have that capability, but it's being really intentional with like what is changing and that human in the loop. So in terms of stack ranking, I think you're putting those inputs in. So you're essentially making the decision at that gate. You're setting the gates. So as long as a human is setting the gates and we feel like we are making that decision internally of like what would qualify somebody in versus what would qualify somebody out, like based on what would be in a resume. I think it's just like not getting carried away, obviously using good judgment when you're putting those gates in place. Like that's been happening forever. Totally. So I don't think that there's anything dramatically different about So it's not being so...

I'm not afraid to run fast as you know. So it's like still like running in that direction, but obviously like having an air of caution and really like being intentional to like what is changing? What is different about this approach versus what we have done historically? And uncovering like are there any legal implications that we should be mindful of there? It's also if you're using tools and technology, really trusting your partners. If you're not building something yourself and you're using a third party vendor, those relationships to me are so critical. Like I get in the thick of it with any vendor that I'm working with and really like establishing those tight relationships to understand their safety nets that they have in place as well and making sure we're clear on that.

And then obviously with our legal team reviewing anything that we're doing, and we have a great internal legal person and she's all over it, in a way that will slow me down if that is required. Yeah. Okay. Let's... I wanna fast forward to 2030. Right now it's 2026, towards the end of it. 2030 is not that far away, but we have seen how crazy the advancements have been since even we worked together a couple years ago. From the... Let's let's just focus on the people space. Okay? And actually, I... I'd be open to to to you kind of extrapolating to, like... Because you're kind of across the whole business now. But, like, where do you think things will be in three years from now from an AI perspective, especially in the kind of people and maybe even just, like, people not just, like, recruiting, but, like, across the business?

Like, where do you think we're going? Oh, jeez. I think it's dramatically different than it is today. I mean, I think the recruiting process, similar to a sales process, becomes heavily agentic. I think... Like, right now, the feedback that I'm hearing out there is that people don't like interviewing with an agent, right? And it's kind of kind of like too bad, right? Like, think that is going to happen. So I think you get to a point where recruiting shifts dramatically different, but there is somebody that's still responsible for those agents. So it's kind of what you said. I think you have your core recruiter actually becomes a manager of agents and they are the ones that are making those decisions. But that whole interview process, like the days of doing 10 back to back phone screens are gone.

Right? And the nights of like, I spent a lot of nights watching reality TV and sourcing earlier in my day. Like those nights are over, right? Like I think you really see people leaning into much more strategic work and even like the performance review process. We used AI this time and it was the easiest, like I... It was an moment for me of like, this process usually sucks. People hate it. And I did not get that feedback. It was so easy for people. How do they do it? Did they did they use like a clause skill and provide it through there? So, we partnered with Lattice. So Lattice has AI functionality and it's... Again, it's getting people over the hurdle of like now you are, you're recording more information. Right? So that's a change management and a transformation that we had to go through to get people comfortable with putting in enough information that it could access it.

But once you get over that, it's drafting these performance reviews that are based off of factual data. So your starting point, your ability to then do performance reviews, like it's like that, right? And then you're responsible for the output, you're responsible for the quality, and we want to make sure that people are obviously giving it the time and attention that it does deserve, but it's not that arduous process that it was before. And that was amazing. So I think you start to see that. And then the real life, like the coaching, that in the moment coaching through agents, I think for various roles. So like your ability, like you just won't need a team of HR VPs in that way. Like I think an HR VP, when I think about 2,030 at my organization, normally I think, Okay, headcount's growing, so that team has to grow.

And I don't think that's the case. It's more so of like setting those agents underneath an HRBP that can do that level of training and coaching in the moment for employees as something is happening. And then obviously, like there's legal implications. We're dealing with people. So that's where that human in the loop really stays in place. But I think like similar to any other organizations, what we're doing now is mapping out like the domains of the people function, putting all of those use cases underneath. Where do agents fit? Where do people sit? And where are, like, the stewards of those agents? So that is, like, where I'm headed anyway. And I think a lot of other organizations are also moving in that direction. Yeah. You said something something interesting there. So... Okay. You've got if if we get to a point where people are interviewing with a agent.

Right? Because I'm assuming you're already seeing what we're seeing, which is a lot more applications coming in. Yes. Because people are just using agents to create an application that is completely bespoke to the job, what the job wants. So it builds their resume based on that. We already see companies like Heygen and others are getting really... Hagstrom, that are getting really good at video. I went on a vacation, and you you know my wife, Jen, and she she knows that I can't stop working on vacation. So I I promised myself I was only gonna do the fun work, the stuff that I I I wanted to do, I've wanted to do for a long time. And one of those was to make a a little commercial for us, a little TV spot that, like, would be about the the whole idea of in the nineteen nineties, early nineteen nineties, there were companies that, like, didn't wanna adopt the Internet.

And it was the whole idea was, remember the companies that were slow to get online? Neither do we. And I was like, okay. Like, that'd be a fun thing to, like, show, like, people doing white out and, like, having, like, typewriters and that kind of thing and then being like, we're a well oiled machine. So I go on one of these platforms, and I start, you know, doing, hey, this person sitting in the middle of the table, of a conference table. Well, the first time we did it, it put the guy in the middle of the table, like, with the table through. He's sitting through the middle of the wood. But, man, once you refine it a little bit, it got really good. I'm like, this this looks like we went and shot this. So I guess my question is, like, do we think... Can we imagine a world where it can't be too far off, where you have an AI double that it can speak just like you, it looks like you, it's probably hard to tell, especially through a Zoom or virtual meeting, whether it's AI or not.

Are we gonna get to a world by then where the interviews are happening by AI to a potentially AI avatar? Oh, totally. You know, I mean, like, how do we... Once we get to that point, once the tech is that good, like, what do we do? Like, is it in person interviews only? It's crazy. I think in person for when you need a human in the loop, I think in person comes back. Yeah. You know, the same way that like offices are coming back. Yeah. I'm back in an office Me too. Very happily so. I know. I love it. Yeah. It's great. And so I think it's like, it's gonna be a combination where I think like, I do think that like everyone's setting up their second brains, myself included. And so like once that second brain gets so good and the personalization of that, and I can get the slop to not be slopped to actually be the way that I would write very specifically, which like again, everyone's putting all of this effort in.

I think you could see that. Like people are building their second brains for work, but also individually. And so if you're able to build that, like I think you could because the interview process is painful. So, and I don't think that's bad. I would love that. If you could have two avatars and they figure out if there's a good enough skill set that would warrant me to have an in person interview, because if we're actually going to be two people that are working together, I think that's important. I would happily like put somebody on a flight and have that dinner and decide if it's a good fit. Yeah. But if we're all going to have agents working under us, like why wouldn't we let them handle that painful part of the process anyway? Like which... Where it's like where people are potentially wasting time. Yeah. So I would do it. I saw a movie recently where it was illegal to have you can actually we got to a point where you could have an AI double, a body double, a physical body double.

You can't even tell. It was illegal to have it. And these two people were using the AI doubles to date. Yeah. To, like... And to... Because they didn't wanna be... And I can just imagine, like, how... When we get to that point where we can use an AI double to be talking for us and everything, like, what's gonna happen to the interpersonal stuff? And we're already seeing, like, I think some... Like, some of these people that are coming into the workforce that were, like, in COVID during, like, key Yeah. Times in their development, like, they're... They they don't have, I don't think, some of the interpersonal skills that I think a lot of other people got by, like, being... I wouldn't... You and I, like, we were probably in office for the first fifteen, twenty years of our career. Right. You know? Like, there wasn't an option.

You had to be interacting with people. Right. No. I think it's it's terrifying. Well, it's super exciting for me professionally. As a parent, it's terrifying because I have high school age kids, right? And so I think about this all the time as to like, oh, I don't... Normally I'd be like, go sling software somewhere or be a software engineer or wherever. But it's changing so rapidly so that it's really hard to think about how do you ensure that that next generation is socialized. Actually, kind of going back to work for a moment, one of the things that I think we're really going to have to figure out is we talk about stewardship. To be a great steward and to manage the context, need domain expertise. How are we passing that down?

Like that I'm genuinely worried about because if you talk about like collapsing up that bottom layer of an org chart, right, or things like that, like you're not giving people the opportunity to develop domain expertise, which we need. So once all of our stewards of today age out of their careers, what happens then? And that's something that I think we need to be really intentional about is how are we going to bring up the next generation of knowledge workers, which we will need that. We're going to need that, that domain expertise. You don't get it in university. I learned so much more once I started working and getting into the thick of it, as I think everybody else does as well. So that's something that I think needs real thought and attention when I think about the future of work.

Like agents aside, they're all going to be there. But how are you going have enough domain expertise to have the stewardship so that you can drive accountability at the human layer. To add on to that, if you don't... Like, people have asked all the time to us, like, hey, do you think college is gonna be a thing with how fast AI is is moving? We're like, honestly, we're not sure. The thing that scares me the most about about not... About people not going to college is not that they won't learn something, because I can't remember a thing I learned, like, from classes there. Right. But I did learn how to present better. I learned how to shape a presentation and to understand what... Okay, these are the instructions. This is the teachers, what they're looking for.

I understand how to figure out how to get that, and time management of myself. I learned those things through college, and that has been really helpful. And I can't imagine going from high school to a job. Right. Right? And then and then learning how to manage your time when you're, what, 18 years old. Yeah. Well, it's like it's the maturity curve. Right? I think college, that is that experience of learning how to be an adult. And I don't think college is the only route, I think between college and the military are the two often that we see. And that's where people really learn that self discipline that you're talking about and their ability to develop those communication skills and live independently. And so it's like what replaces that? Yeah, and that's where the parenting side of me, I'm like, oh, wait, I thought I really knew the path that we were on.

And now it's a little bit gray, which is, again, super exciting. Lot of things. Nerve wracking when your kids are out. Like, we're not there. My son's graduating in 2030, which we're talking about. I'm like, oh, God. Like, where where's work gonna be? Where's he gonna be? Yeah. So it's it's a lot to take in. Yeah. Well, Lord, thank you so much for coming in. Yeah. Appreciate it. Great talking to you. This is a pleasure. Awesome. As always. Okay. Thanks. Awesome. Thanks, Keith.