Episode Transcript
[00:00:00] Good morning, everyone. We just finished our fifth annual Irresistible Conference, and it was really spectacular. For those of you that were there, I'd love to get you to write about it or tell people about it. But basically what we did for three days, we had about 450 senior HR people, a few vendors, not very many. And we went through day one, which was Imagine the Future, Day two, which was Implement the future.
[00:00:25] In day three, what tools do we need to get to the Future? We introduced HR 2030. I went through the narrative behind it, which I'll be doing more frequently at shows. You guys can see this. We introduced our new certification program. We talked a lot about industry solutions, frontline work, many of the things going on in the airline industry, retail music, had many interesting presentations from many industry HR leaders, and talked, of course, about AI and the complexities of AI. And there were lots and lots of meetings and lots, lots of events. And I want to give you a couple of reflections on what I think I learned this week. A couple of things. First of all, I think, generally speaking, from the companies that were there, the AI issues and technologies are rapidly reaching the levels of awareness and expertise in hr. People are very familiar with what this is about. This isn't black magic anymore, and reality is setting in. So, you know, the funny thing about this technology is it sounds magic if you listen to a podcast, but if you actually use it, not building code quickly doesn't mean you've built an application or a solution. You're. You've just written code. In fact, I have this sort of vision in my head that clog code is going to create a mess because there'll be all this code out there and nobody knows what it does or knows how to use it because it's the structure or infrastructure or harness around it that makes that adds value. So I think people are getting that. One of the discussions that I had had to do with the role of data. And, you know, historically, what we've always done in HR is we've thought about our big technologies as systems of record, and we need to clean it up and we need to organize it and we need to create data analytics tools around it. That's a real issue in AI. It's a different set of problems to solve because the richness of the data is so literally orders of magnitude richer than it was in the past. And we talked a lot about how to manage this massive amount of data, and I think people are aware of that. So I think the market is caught up to the vendors, to be honest, no More selling stuff that doesn't exist. No more making up words that nobody understands. The people at our conference at least are aware enough of the space that they're asking the right questions. The L and D teams are now aware and comfortable and learning about dynamic enablement and dynamic content. The recruiting organizations are very aware of what tools like Paradox or Eightfold or other ones do. And the negative implications of that were positive. And I think the IT teams that are working on data management and quality are worrying about that effectively. And I think one of the most interesting sessions of all, which I'm going to have a podcast on, was with our friends at Lockheed Martin. And I won't give you too much of the story, but Lockheed Martin is a very large company with hundreds of projects that are very technically important and difficult and complex. And they have taken a real enterprise, I mean, I guess I would say traditional IT approach to AI, and they have 250 or more agents now running at Lockheed Martin, and they're managed and they're controlled and they're governed and they're getting value out of of them. So the story there is go slow, to go fast. Don't just go nuts building things and hoping that they're going to work. The experimentation phase, I think, is over so we can move into much more of an enterprise view of how to use all this technology. Okay, now I want to talk about a different topic that came up a lot, and that is the token economics of AI. For the first couple of years of this stuff, we were subsidized in a large degree by investors who threw money literally as fast as they could at these companies to try to build up their infrastructure. And there was no need to look at profitability or return on investment from the vendors. And so this concept of token maxing or use as many tokens as you can was kind of normal. Well, that is ending effectively now because there's real cost to this. When OpenAI and Anthropic go public within an hour, people are going to look at the financials and talk about profitability and costs and revenue and so forth. So we're going to have to pay for this stuff. And at the end of the conference, one of the meetings I was in with a large company, we're talking about the cost of their HCM platform, and I won't mention the vendor name. They're a CPG company, so they're a good sized company, but they're not growing at the rate of a software company. So they have, you know, real budgets for HR tech And they can't just spend money on anything they want. And they had their sales team from this vendor in, and the sales team was talking about the usage based pricing model that's coming to their tools. And by the way, we've been having this conversation with HubSpot too. So they asked the sales team, well, show us some of the transactions that'll be new and how much they'll cost. And they showed them some things like booking travel or scheduling a meeting or creating a performance appraisal or booking time off in your scheduling system in the HCM platform. And it was a dollar for this, 75 cents for that, $2 for this. And they did the math in their heads and they said, this is insane. There's no way we're going to be able to afford this. If it costs each employee a dollar to schedule a meeting or to book their vacation time. And we start adding that up by the number of employees and the number of vacation days and the number of days per year, we just doubled the cost of our system. And they said, the sales team was shocked. And the sales team said, oh, we never, we shouldn't have done this, we shouldn't have shown you this. But that's, this is, this is really happening. And all these stories that have come out in the press about software engineers blowing their budget on tokens, this is a big deal. Now I wrote about this in the Imperatives late last year that one of the big things that's happening in 2026 and here we are in June and it's happening now, is we have to think about the economics of these projects. If the project you're working on is a vanity project and you don't have a big use case that is going to improve performance, productivity or growth, then maybe you shouldn't do it because it's going to be too expensive. What I talked about in my keynote and I'll be doing this keynote more often, so those of you that weren't there will be able to hear it.
[00:06:33] In our particular area of business, in hr, cutting the cost of HR is a tiny return on investment. It's not that high. Nobody cares. I mean, they care to a degree if we can make HR more efficient. But that's really not the purpose of the company. The purpose of the company is to grow, to reach customers better, to add more value in our products and services.
[00:06:54] So the ROI of these various AI projects is not cutting the size of the training department or cutting the size of the recruiting department. It's something much bigger than that. It's more likely improving time to market, increasing revenue, improving productivity of engineering, improving productivity of production or service delivery or healthcare, or improving retention. Things that have huge business economic benefits, not internal economic benefits. So a lot of what I think we need to do in this world of AI is reframe what we're here to do and why we're here. We exist. And what I talked about in the keynote is my new phrase for HR is why we're here is dynamic enablement for growth, dynamically enabling the human capital part of the company to facilitate business growth and personal growth for the individuals. And that's it. And if you sort of think about things that way, and you're a retailer like McDonald's or Chipotle, we had Chipotle on the stage talking about how many more burritos they can sell because they can recruit faster with AI. That's the way you have to think about these projects because they're not going to be free. Now, as I was sort of mulling over this conversation that happened near the end, I was thinking that maybe an analogy of this that everybody could rate to is imagine you got on a plane, which we all do, and you. They said, it's a hundred dollars to get on the plane to go from San Francisco to Miami. And you got on, you picked a seat, and somehow they told you what seat to get, even though every seat now seems to have a different price, and which is a weird phenomenon. And you got off the plane, and they said, okay, everybody, before you get off, take out your credit card and we will charge you on the way out of the plane the cost of the flight. And then you're going to say to yourself, what's the cost? And they're going to say, we can't tell you that until you get off because we don't know. The fuel prices could raise. There could be a change in demand, and at the end of the flight, you'll get a bill. And you know what you would do? You wouldn't take that flight, you wouldn't use that airline. It's a little bit like getting into the Uber, not really knowing how much the car is going to cost you to go from place to place. But they do give you a good estimate. We don't get estimates on tokens. Nobody can estimate how much something's going to cost because I don't think even, even the vendors know, because the cost of the token and how efficient the model is and how well you're using it or not using it, it's very Hard to predict that. So the bottom line to me on the economics of these tokens is if we're going to go to usage based pricing, which I guess we are, although if you think about the consumer market, usage based pricing led to monthly pricing. We had usage based pricing in the phone industry for a long time. You know, long distance calls were expensive and so forth. And then they finally said, once we had enough capacity, let's just have a monthly charge and you can use as much as you want. So, you know, that'll probably happen here, but we're not at that point. So if we're going to have usage based cost, we're going to have to think about the value of the project and the complexity of the project and the design of the project. So this idea that somebody's going to be whipping up Claude code and just spending all night generating stuff and then the next morning we're going to use it, that's not going to happen. I mean, not for very long. Because that thing that this person created might have cost $3,000 just to create it and then to run it. It might be, you know, a hundred bucks every time somebody runs it. I don't know, it depends on what it is. But different, different way of thinking about this. So where we're going to me in enterprise AI and HR is back to a world we've been in before, which is enterprise applications, enterprise solutions. It's not that different from my days in the mainframe world where you had to cost justify why we were going to spend $10 million on a new mainframe. And we knew that over time we needed the capacity, but we needed projects to cost justify it. This is the same thing here. And of course in the middle of this sort of economic issue, the products are very immature. So there's a lot of agony about which model, which vendor, which tool do we use, the incumbent vendors, do we buy a new vendor and so forth. And I would say another trend that I picked up at the conference is one that I think will make the enterprise software companies very happy. Most of the customers that we talk to want to use the tools from the vendors they have. They're happy to use Microsoft if they have it, because it's already there. They don't want to rush out and replace it. And I'm not saying that there aren't great things from OpenAI and Anthropic. There are, but not everybody wants to jettison what they have. And the incumbent technologies and data and utility that they, that, that we have in our companies is extremely valuable. So the future of HR 2030 is not a bunch of rip and replacement by any means. It's rethinking what you're trying to do in this new world of dynamic enablement for growth and taking the technologies that are available from either the vendors you're working with or somebody else and building those kinds of solutions in partnership with it. And one of the things that came up in all of the case studies and all of the roundtables is we can't do this without it because these are IT laden issues. And in the Lockheed Martin session we talked about this a lot, that until it could come up with a framework for enabling all of the business functions, not just HR, by the way, to build and manage and administer and monitor different AI solutions, it was chaos. And they can't operate that way. They're a very, very important contractor. And defense companies, they can't operate with chaos. So I don't think any of us want to operate with chaos. So anyway, all that came up, the other thing that I just want to reflect on, and I talked about this and actually kind of broke out into tears a little bit at my speech, is the amount of goodwill and warm and collaboration and real affinity HR people have for each other. And I really appreciate this as an analyst. This is an industry of people who really care about their companies, their employees, their teams, their careers, of course, their skills, their quality of work everywhere. I feel fortunate myself, and I think most of you probably do, to have landed in this domain where there isn't a lot of politics, there isn't a lot of pride, there isn't a lot of blustering in hr. It's a lot of soul searching, problem solving and thinking about people issues. As you'll see in the book we're launching in the fall. We did a preview of the book and we gave everybody a little bit of a piece of the book to read. The idea of human capital is going to be just as important in the world of AI as ever before. And it's funny, you know, We've got the SpaceX IPO today, there's a bunch of funny stories about robots, robot dogs. I don't know that we can stop robots, robot vacuum cleaners, robot cleaning tools. I mean, one of the companies that was there actually was one of the largest providers of cleaning services and facility services for companies. And my CIO told me that they absolutely do believe that a lot of their workforce will be augmented by robots very soon, you know, pretty quick. But I think what came out of the conference for me is amongst our our peers. The human capital issues are not going away at all, at all. So these companies that grandiosely claim they're going to replace a bunch of people with a bunch of agents, they're just not seeing reality yet. They will, they will over time. All of the technology evolutions tend to start with fear and then later become much more pragmatic and we're kind of getting to that point now. I also noticed that OpenAI has quieted down about the number of jobs that they think are going to be eliminated. By the way, I would recommend that you do not listen to economists who work for tech companies. They're very self serving and I think a lot of their findings are very biased. Look at real economists, look at micro economists. In some ways I sometimes think that we're like a microeconomic company because a lot of what we do is talk to individual companies all day so we see the real feet on the ground stuff. So I want to thank our team who did it just in spectacular job of running the conference. The USC team who we have a very deep admiration for and we work with them now for five years, is the fifth year of the conference. All the HR professionals and leaders who came and participated and joined us and we'll be. There's a lot of new stuff coming from us over the next couple of months here that you guys are going to hear about soon. We're very close to launching Galileo on the Microsoft Copilot. That'll be in a few weeks. So those of you that are Microsoft shops, you're going to be able to get all this stuff in your Microsoft infrastructure. And one of the cool things about Galileo in the Copilot is it's going to be enterprise scale so everybody in the company will be able to get the benefits of these modeling tools. Okay, I think I'll cut it off here and save the advertising for later.
[00:15:47] Have a great weekend everybody and happy summer. Summer solstice. We're going to be into the nice long sunny times of the year and let's enjoy the summer while we can because there'll be lots and lots of excitement in the fall. Bye for now.
