Episode Transcript
[00:00:00] Good morning, everyone. Today, as we enter 2026 with many, many, many things to think about, I'm going to start off the discussion of AI architectures. And the reason I'm bringing this up is because it's very clear to me that 2026 is the year of enterprise AI. And what I'm referring to here, as you'll read about in our imperatives report, is not AI as an individual personal assistant to make your job easier, which is certainly going to be explosive, but even bigger than that, AI as a way to redesign and automate and improve our business processes themselves. This is a much broader context and much bigger return on investment. And the argument that I would make to you is that actually the big ROIs of your tens of millions of dollars invested in AI and it is going to be that big, is going to be in the workflow and the business process improvement, not in individual productivity. Because if we don't change the jobs and the responsibilities of the individuals in the company, and we're not going to make that much progress on re engineering and business performance improvement with AI.
[00:01:15] So how do you do this and what do you consider in the process? Well, it turns out I've been around long enough that this is all familiar to me because we went through this architectural transformation in the early 2000s in the Worlds of talent management.
[00:01:30] And so let me give you sort of the simplistic story to help you think this through. We have studied HR now for close to three decades and there's been a lot of change. But perhaps the biggest change of all is the explosion and increase in the number of topics and processes that we have to discuss and work on. Right now we, we have almost 100, around 98 individual, what we call capabilities in HR. I know that sounds ridiculously large, but when you see them, you'll agree with me. They're all things that we have to do to run our companies. Well, some of them are management things, some of them are pure HR things, some of them are compliance things, some of them are regulatory things.
[00:02:18] But they all have to get done somewhere. And the reason they get done in HR is because they tend to be things like recruiting, like skilling, like workforce management, like promotion, like leadership development, like coaching, like performance management, like pay equity that individual managers can't really do themselves very well. So we've got these 94, 95 capabilities in HR that traditionally have been organized into centers of expertise or centers of excellence built around a service delivery model. That's the traditional level 2 HR function. If you're In a small company, this is all being done by a few people. But the bigger you get, the more you end up building these specialty teams that do these specialty things. And then it's stitched together with business partners, HR support teams, call centers and, and as much technology as you can find. And the ideal vision, of course, from the HR technology marketplace is that all this would be automated. So before AI, we went out and we bought individual talent management or HR tools from individual vendors. And 25 years ago there were hundreds and hundreds of vendors selling recruiting tools, learning tools, pay analysis tools, analytics tools, payroll tools, workforce scheduling schools tools, shift management tools, tools to do compliance, et cetera, benefits, on and on. And then the big human capital management vendors during the times of the cloud said, well, wait a minute, we'll do all of that for you. And they bought up or built out more integrated systems. And these days you can get an integrated HCM either as a mid market company or a large company. And this goes from everything from Hibobi to Lattice to Workday to UKG to SAP to Oracle, et cetera, that does almost everything in some fashion with payroll. And it feels a lot easier. Of course the problem with this is these systems are not super intelligent, they're not predictive, they're not very good at analytics, and every company has to implement them differently. So even if you buy a great system, you're spending a lot of time configuring it and customizing it for your particular company. Okay? That's the state of the market. Along comes AI and all of a sudden everything can be done differently and better. We can interview people electronically and assess their skills using AI agents or AI interviewers or AI sourcing tools or AI based testing tools. And you can extend that to every other part of those 94 things.
[00:05:01] So what we did is we've been watching the market now and working on it. Of course with Galileo is we now have In a sense, 94 or more potential AI agents that you could implement or buy or build in your company. So the large vendors are building these agents for you little by little, and they each have agent roadmaps. And the eventual direction is that these agents will appear. You'll have to pay for them, I'm sure, in your HCM systems, and you'll have a whole bunch of new agent features in your HR technology, but they're not doing it very fast. So as along with the $3 trillion invested in data centers, probably billions, I don't know, the number is being invested in AI HR startups. And so There are literally hundreds of really amazing, very creative, innovative AI based tools entering the market right now. Many of them are quite proven and we know where most of them are to do things that you cannot do in your hcm. And the HCM vendors cannot possibly build all of these AI use cases. They don't have the time, the resources or even the imagination.
[00:06:12] So you're going to be sitting here during 2026 talking to your existing incumbent payroll HCM vendors and then saying, well, what if we bought this, what if we bought that, what if we plug this in? And I'm sure the vendors are going to have marketplaces and integration frameworks to help you find the partners that you need. So presumably these things will all work together. However, this is a little bit more complicated than it looks. And as I'm going to explain to you in the webinar I'm doing this week for clients and next week for the rest of you, it doesn't work quite the same way it did with hcm, because each agent has its own data system, its own levels of intelligence in its own essentially agency to make decisions. And unless the agents are talking to each other, which there really isn't a lot of protocol to do yet, a decision made on pay by a recruiting agent may need to check in with a decision made on pay benchmarking by the pay equity agent, which might need to talk to the DEI agent about the diversity metrics relative to pay, which might need to talk to an agent that looks at competitors pay for various critical roles, which might talk to the performance management agent, which is looking at pay relative to performance across the company. You can see where I'm going is that systemic hr, which is the direction I know everybody's going, isn't really going to be possible if you buy 100 agents or more from independent vendors or build them yourself because they're not going to work together.
[00:07:51] So we're going to have to figure out how to get these agents to work together. Theoretically, from an IT standpoint, that could be done with APIs. And I've talked to the CTO of Workday and SAP and others about this, and that's their vision, is that there will be APIs. And so all these things can talk to each other, but that means you're going to have to program them and explicitly tell them what to ask each other when. And that's complicated in and of itself, when a lot of them aren't even defined or don't even exist yet.
[00:08:20] So we're in the beginning of building a Whole new architecture for our human capital processes, but we don't have all the pieces. So how do we do this?
[00:08:30] Well, two things have explicitly, functionally changed here, or really three. There's the issue of the data and where it's stored and how it interacts with. There's the issue of the agents talking to each other and how they talk to each other, not technically, but what they say and what decisions are made, where. And then there's the issue of the management infrastructure and the orchestration, as the word is called, of these intelligent systems. And the way we see it, having talked to many, many of you over the last two years about this, is we're going to be having super agents that do the work of more integrated decision making processes across these 94, 95, a hundred or more agent potential opportunities. And already this is happening. This isn't something that's going to happen. This is happening. If you look at the existing product roadmaps of most of the HCM vendors, including the big ones and the small ones, they're already introducing small agents to do all sorts of things. And I don't mean just writing job descriptions and sending emails to job candidates. I mean much more complicated things. Sourcing, assessment, leadership development, career planning, skills development, et cetera. And by the way, there's a new wrinkle here, which is that much of the human part of hr, where we would interview people or do needs analysis or do consulting, can also be done by agents. Because there will be, and there already are agents that can survey people or ask people questions or assess their needs from their interviews or from their meetings or from their discussions or from their inquiries to give you input into these agents. So there's probably another 20 or 30 agents that will be created just for gaining information or listening or asking questions about the stuff going on in your company.
[00:10:31] If there's a harassment claim or a bunch of people are showing up late, or a bunch of people quit, there will be agents to investigate these anomalies and get data back into the system so that the system can make better decisions and inform us as humans what we need to do.
[00:10:47] So this new architecture, as we call it, is going to be a family of agents and super agents working together to implement the human capital practices that we hold dear. And not everybody is going to optimize everything, because not everything matters to everyone. If you're an airline and you can't find enough flight attendants, you're going to focus on that.
[00:11:12] If you're a retailer with high turnover and dealing with massive changes in volume because of, say, a weather situation or a change in the market, you're going to be dealing with that. If you're a hospital dealing with a bunch of nurses, they can't come in because there's a fire or some other catastrophe that creates demand, you're going to be dealing with that. If you're a software engineer and you're trying to hire a 10x engineer or a 10x scientist, if you happen to be in the pharma industry, you can have a completely different set of issues. So these agents, intelligent that they are, are going to have to be tweaked and configured and tuned for each company.
[00:11:51] Now, it turns out we have been thinking about this a lot and one of the things I'm going to discuss briefly in the imperatives is our new AI HR blueprint built on the systemic HR framework. And this is an offering that we are now talking to clients about. And you can each talk to us about it. We're not putting it into the public domain, at least no time soon. But you'll see that what we've done is group these agents into autonomous workflows that fit together logically. Now, this is a great way to go to market and think about the massive number of new tools coming to you because the individual vendors can't build everything.
[00:12:32] So you as a buyer are going to have to make decisions as to how much automation you want to implement in what areas. Obviously, the critical areas of opportunity are in employee services, in recruiting, in learning and development, and in the HR business partner role. And I'll explain why we picked those four areas first because, well, there's reasons for it. When I do these webinars the next couple of weeks here, where is this all going? Well, where this is going is you're going to have. Oh, and there's one more thing. Given all this and these wonderful opportunities to rethink how we do HR and how we do HR tech, there's one final wrinkle. AI is a tool set for building things.
[00:13:16] You can literally take Galileo and Claude code or any of the agents in the market. I would say Claude is ahead and you can literally tell it in English or whatever language you speak, what you want the agent to do. And it will build in code an agent that does what you want it to do. Just as an experiment over the holiday, I put together a very simple, less than one page description of a performance management agent, which I called Galileo Performance Management just for kicks. And it built for me in about half an hour, an entire application for Performance management, which was about as good as many of the tools companies used to sell in the early days of online performance management. Now, it didn't build interfaces to everything else. And I didn't try to explain to it all of the if this, then that, ask this, ask that, because I didn't have the time. But we could literally sit down with a tool set from the public domain LLM market, or Galileo, or a tool set that you might get from your HCM vendor, and you could build these agents yourself. So rather than waiting for a vendor to build them for you, you're going to be either building them or changing them or complimenting them. And in fact, I really think the vendors are going to have to think hard about how open they make their agents. Some of them might even be put into the public domain so that you can buy them as source code or as prompting code and rebuild them in your experience. And that's not a business model these companies are used to. They're not used to giving away their code. But I think we're going to have to deal with that issue because it is extremely easy to build things with AI. And remember, we're only in the first or second or third year of this technology. By the way, the building or citizen developer market for AI is expected to be 20 to 25 billion dollars in itself. So the HCM or HR technology vendors, each of whom are trying to find revenue opportunities in this new space, may end up building development tools for you or offering you development tools instead of completely packaged applications, because it is so easy to customize these things. Remember in the old days, if you bought a piece of software and wanted to customize it, you probably couldn't. Or you hired an engineer or a consulting firm and paid millions of dollars to have it tweaked and customized for your needs, keeping your IT department very busy trying to figure out how it worked. Well, that problem is going to be way easier than it ever was before.
[00:16:01] So you have to incorporate into the architecture the idea that you as an IT or HR professional will also be tweaking these things to get them to do some of the use cases that are unique to your company. Now, there's kind of two angles to all of this. One is, as a buyer and architect and leader in hr, you want to think holistically about what you license. Because if you don't think about the super agent scenarios, you'll end up with a bunch of independent agents that don't work together from different vendors, with different levels of quality, with different levels of customization and different levels of integration, which would lead you to want to do business with the bigger vendors. But the bigger vendors are slower, they're more complex, and they're more expensive.
[00:16:51] So the first issue is, where do you start, what do you buy, and what is your big framework picture? And that's where I think working with us will help you a lot, because we've been down this learning curve now for almost two years, and I think we can help you get yourself prepared for the marketplace. The second is to get smart about what's going on so that when you talk to these vendors, and I don't mean talking to a salesperson, I mean, you're probably going to have to talk to product people or senior people, ask them where they're trying to go with these new agents they're building, what are the bigger use cases they're trying to solve, and how sophisticated is their thinking relative to your needs? I mean, one of the best examples of this to me is the work that UKG has done in workforce management. And I really am excited about some of the stuff they're doing. UKG is by far the leader in frontline workforce management of 80,000 companies running various forms of Kronos or UKG HCM to optimize shift management, scheduling, work assignments, skills development, and other issues we face in frontline work. And it's enormously complex and different by different industries. If you're an oil company, if you're an energy company, if you're a transportation company, entertainment company, healthcare company, gym, whatever, the scheduling issues and the scheduling requirements are different. Well, rather than just giving everybody an automated tool for scheduling, which they kind of already had, they went beyond that and they built a system they called dynamic labor management or dynamic workforce management. And what it does, because it's an integrated hcm, is it looks at the shift schedules of all the people in a given geography or given location, and then it looks at demand signals and it makes recommendations to the workforce manager, not necessarily the HR department, whoever might be running the group, and says, you know, you're going to need more people on Thursday because the demand seems to be peaking and you're going to need people of these skills in these locations and these hours. Here's the body of people you have. Would you like me to open some racks to hire more people? Would you like me to change the pay so people will come in on extra hours for overtime? Would you like me to change the shifts or move people from place to place to get the critical skills into the right Demand locations. It does that for you. Now I'm not saying it's perfect, it's probably still pretty new, but that's an idea to me of a super agent that would not have existed had these individual agents not been conceived first. So we're going to be in this really fascinating world of you either discovering or exploring or inventing, in a sense, these super agents and how you want the individual agents to work together. And I know some of you are going to want to wait for the vendors to do this, and that's fine. And eventually most vendors will get to this point. But remember, as with everything else in HR that I've been observing for the last three decades, the stuff keeps changing. There will be new ideas. I mean, one of the big topics in the imperatives is talent density. So let's suppose you read all our research on talent density and you rethink the way you hire and you don't go out and open recs and just hire every time you need capacity anymore. I mean, you may have to if you're in a high volume situation, but. And you say we're going to reskill people, or we're going to move people, or we're going to change assignments for people to fill demand, to increase talent density. Well, that may not be embedded in the ATS agents you bought from your recruiting vendor. They may not have even considered that a use case. And there will be hundreds others that will come along.
[00:20:44] So this is going to be a new, innovative, creative stage in our transformations. And I think 2026 and probably 2027 and 2028 are going to be enormous years of HR transformation. And I don't mean transforming HR as an HR department.
[00:21:03] I mean transforming the way we think about human capital, the way we think about leadership, the way we think about teams and structure, the way we think about performance, the way we think about jobs and skills and roles with a much more super worker approach.
[00:21:23] Because while we're doing all this business process thinking, every individual job is being completely superpowered by AI analysts, recruiters, business partners, everybody L and D people. We're all getting super powered. So these super agents, as I call them, are going to be super powering every one of us.
[00:21:48] And over time, as most of you know who involved in technology, this gives us autonomy, this gives us freedom, this gives us creativity and opportunities to add value like never before. And I do believe, and you'll see this in the research, in the blueprint, that we will end up with HR departments that might be 30% smaller than they are today.
[00:22:11] That doesn't mean these people are going to be laid off. It means they're going to be doing different things in the company or in hr. But that's where this is going with the ultimate goal of not just reducing the size of the HR department. The ultimate goal zoom proof, speed scale, quality of employee experience, skills, capabilities and business performance and business productivity and customer experience.
[00:22:38] Okay, very complicated topic. I tried to make it simple. Come to the webinar this week for clients, the following week for the rest of you and all of the information I'm talking about will be in Galileo starting next week. So if you don't have Galileo, get it because you'll get access to all of this and any of these topics that appear confusing to you. You can literally talk to Galileo and Galileo will explain them to you, teach you about them and apply them to your company's issues. That's it for now. Bye everybody.
