Episode Transcript
[00:00:00] Hey everybody, I want to talk about HR 2030 a bit and explain it to you because there's going to be quite a few announcements coming about what we've done.
[00:00:11] So what we did is quite extensive and you won't see all of it because of some of it's proprietary. But we basically took all of the research we've done in the human resources profession, the 95 capabilities of HR, all the different use cases, 221 sort of business applications of talent and HR practices, and built out the best we can, an agent architecture that would run a highly dynamic, highly growth oriented, near real time HR function using AgentIQ technology.
[00:00:52] And we built it around an architecture of a three tier architecture.
[00:00:57] Super agents, agents and personal agents. A personal agent is like Grokbot or Muse or this new thing that Microsoft's working on that sits on your desktop or the Copilot. An agent is an action oriented AI that does things that has guardrails around it, it has specific responsibilities, it can't do anything. We have to give it to very specific responsibilities. So it's really good at a few things, not everything. And then a super agent is a coordinating agent that makes business decisions and optimizes the company around the 150 or so agents. And then the agents themselves have three functional types. Monitoring agents, which you might call ambient agents that monitor things, action agents that do things, obviously making decisions in the process. And then rules agents that store business rules and culture and behavioral rules and other things to inform agents about what they should or shouldn't do. And then there's a fourth type of agent we created called a data agent that deals with the messy conglomeration and integration of data to make it easier for all the other agents to get the data they need so they don't have to do it themselves. And that could be put into another layer too.
[00:02:20] As we went through all of the use cases in HR, we ended up coming up with about 150 to 160 of these agents. And what we then did is we built it into a very, very comprehensive database essentially that describes the inputs and the outputs of each agent, the data structures or entities that each agent touches, like a job candidate or a requisition or tax payment or something like that. And then we created a model with a whole bunch of characteristics so you could manipulate the agents and look at them in different directions, including the business process that the agent does. In other words, there's a process flow mapped against Oracle, SAP workday, the data elements that the agent touches and the various 220 business use cases that it could be used in. And believe it or not, it all hangs together and works really, really well. I mean, somebody could probably take this architecture and build an agentix system from the architecture because there's a lot of intellectual property in the architecture itself. But that's not really why we built it. The reason we really built it is to help companies plan their strategies. Because if you go out there and buy a tool, a recruiting tool, or a sourcing tool, or an AI interviewer, or an AI based training tool of some kind, or an AI assessment tool, or, you know, there's hundreds of things out there now you're going to ask yourself, where does this thing fit in my long term plans with all of the other things we're going to want to do? And a very significant number of the agents you're going to use are probably going to be coming from your incumbent HR vendors. Despite the SaaS apocalypse idea that was promoted for a year or two that all the SaaS companies were going to vanish, actually the opposite is going to happen. They're going to be fine and they're going to offer agents of their own that will be options for you to use for some of these 151 functional areas. Not everybody will do all of these things because you. There's no reason to implement new technology when the stuff you have is working. I mean, in fact, there's no reason to use AI at all if you don't have a business case for it. We're not implementing AI for its own sake. We're implementing AI so we can do things better, faster and more strategically or more intelligently.
[00:04:55] Now I won't talk about the super agents too much. They're sort of the coordinators and the agents. But let me talk about the personal agents because that's the hot topic in the last couple weeks. Weeks. So if you look at the announcement by Muse Meta's announcement of their personal agent which caused their stock price to go up by 10% or more, I think it was 11%. The tool called Grok Bot by Axe or SpaceX and the potential tool from Microsoft called Scout, which isn't out yet, but it's basically the same idea, which is an application that works with Copilot.
[00:05:31] These are maybe the most powerful AI experiences we're going to have because they're going to be like what we always wanted Siri to be. On your iPhone, if you have an iPhone. Let's just talk to the computer on our phone, on our phone, because it's the one we're carrying around with us all day and tell it what we want to do and ask it to do it for us. And then we have a personal assistant that can read our emails, book a flight, call an Uber, change the temperature in my home air conditioning machine, if I have one, figure out which restaurant to go to, figure out how many calories are in this meal, whatever it may be.
[00:06:10] And all of that stuff is kind of available in different tools on the Internet. But wouldn't it be nice if I could just talk to one system and it would do all of that stuff? It maybe it would call an Uber, maybe it would call a doordash, maybe it would order from Amazon and so forth. And of course that means it's a huge market and everybody's going to want to get in it. Google's going to want to get in it, Microsoft's going to want to get in it, Amazon's going to want to get in it. Obviously Meta, Facebook and X and Apple should be in it, but for some reason they're not. They seem to be behind. Maybe they don't care because all these things run on Apple and they get 30% of the revenue. That idea in the corporate world is giant, giant, giant, giant. Let me just give you sort of some important things to think about here. So if you look at a lot of the new AI tools we've been examining in the world, you know, an AI tool for recruiting, an AI tool that does interviewing, an AI tool that generates instructional training content, an AI tool that schedules your hours if you're a nurse and allows you to reschedule things, an AI tool that lets you get paid in advance and get a payday loan.
[00:07:19] You know, lots of things like that are coming out now from different vendors. Why would I have a vendor provided platform for any one of those? If I could get a personal agent that could do all of that stuff. The personal agent will not only have APIs and plugins to all these other vendor products, which we're already seeing from Grokbot and Muse, but even better than that, they know everything about you. They know where you are, they know what time you got up, they know, they know your emails, they know what documents you've created. If you're connected to Outlook or Google or SharePoint rather, they can infer your skills.
[00:08:02] They probably could have and do have your recorded meetings. So they know your voice, they know your language, they know what you're working on because they can hear from your meetings and also from your emails what's your activity level and what's your activity working on and who you're working with and how well you're doing on your various projects. They know how much money you're making because they're connected to your payroll system. They know your benefits, they know your family stuff if it's in the HR system. Now, a lot of this is personal information and has no relevance to work at all. And we'll have the same standards we do now for data privacy that we have in our corporations already. But this agent can use that information for positive things.
[00:08:46] So when you're late for a meeting or something's going on and you want to get involved in a project and you miss the meeting, or you're behind on a project and you want somebody to help you, or you learn something new and you need to divert the decision you were going to make and change it, or you need to call somebody quickly because there was an update to a product or a service and you want them to know about it, or you want to get your opinion or your data to them, et cetera, et cetera, et cetera. Or I want to get a raise, or I want to get a better job, or I want to know why I've failed at this project, or why did everybody look at me funny at that meeting when I was talking about this new idea I had?
[00:09:24] I mean, whatever it is, this personal agent is going to be able to help you. It's the world's potentially best personal robot that you'll ever have.
[00:09:34] Now, this doesn't really exist in the corporate world yet, but it's going to come in a big hurry. And it may come from Microsoft, but it may come from Meta, it may come from somebody else.
[00:09:44] I mean, we have, I think I've told you this before, we have a digital twin running in our company right now that's indexing, using AI, all of our emails, documents and meetings.
[00:09:56] So, and some of that's in Sana, some of that's in Microsoft, and some of that's in other systems. I can find out anything going on in our company from anybody if they're not there, because I can ask their personal agent what's the status of the proposal to so and so on such and such? How did the meeting go with so and so talking about such and such, et cetera, And I will get an answer within a second or two, even if they're asleep or on vacation. Now, on the positive side, for the employee, if I'm an employee in our company or any other company, And I say to the personal agent that we already have, would you tell me what my strengths and weaknesses are and what areas of the business I should be focused on spending more time either learning or spending my energies on?
[00:10:47] Where would I best apply my skills to add more value to the Josh version company? Or to get promoted, or to get a raise or to close this deal with this client or help this customer solve a problem?
[00:11:00] It'll know. Now, it's not going to be perfect, but it's going to know a lot.
[00:11:04] It's going to know a lot more than my manager, by the way. My manager might have a lot of opinions about a lot of that stuff and certainly will help, but the AI knows in some sense, everything.
[00:11:16] Now, I don't want to scare you, I don't want to make this sound creepy, I don't want to make this sound out of range, but this is going to happen. Let's just be realistic. It's going to happen. So HR 2030 is built on this three tier architecture.
[00:11:31] And the personal agent part of it is relatively nascent so far.
[00:11:37] But imagine now that the talent mobility agent, the employee experience agent, or the skills development agent, or whatever agent you want to talk about, wants to introduce an intervention or a change to the company because of a strategic change the company's making. And we want to train people on something, or we want to inform people on something, or we want to ask for their input, put on something, or we want to give them some critical skills that we know they need right this minute. It will not have to talk to the person, it will talk to the personal agent. And the personal agent will help the super agent or the back office agent give the employee exactly what they need in the way they want it. Because some employees are on the road, some employees don't have computers, some employees are disconnected, some employees are busily working on other things at their desk, et cetera. So this architecture is really, really exciting. And it of course throws a wrench in the works of every software company in human resources because everybody's got to adapt to this. If you're a payroll company, if you're a recruiting company, if you're a training company, if you're providing coaching tools, whatever it may be, all of that stuff's going to have to be built in there. I think a lot of these coaching vendors that are building leadership coaches are going to be squeezed significantly by this because the coaching capabilities of a native generative AI are very, very, very good. Now, without a huge amount of training, Galileo, which is trained on SHL and several other coaching models is extremely good at coaching already.
[00:13:22] I think there will be coaching tools, but they will be impacted by this. I think a lot of the companies that provide advanced scheduling systems, employee experience tools, survey tools, products that give people developmental feedback for well being, all of that stuff will sort of fold into this personal agent area. Now, how do you get your hands on HR 2030? How do you use it? Well, first of all, you kind of have to talk to us because it's a little more complicated than just picking up something and buying it. We have a tool that is a tool we use with clients that will walk you through a process of talking about your business problems and your business strategies and your business priorities to help you decide what part of the AgentIQ architecture you should focus your energies on. And we're spending a lot of time mapping the HR2030 architecture to vendors in the market.
[00:14:17] And we are also starting to scratch our heads on how to do a maturity model on this because of course, a lot of you have a lot of AI tools. If you're a vendor or a software company and you want to see what we've done, we'll be happy to talk you through it. And if you want to engage with us, if you're an AI engineer or an architect or an IT architect working on HR stuff, I think you'll find it very, very helpful to sort out and filter out how the systems architectures are coming together, what the security systems look like, and how the business rules and other capabilities of these agentic architectures work. And if you're a chro, this will really help your team get organized around the flurry or swarms of agents that people are trying to sell you. And that's really our goal here, is really to educate you and assist you in building the right architecture for the future. Which Galileo will be a part of that, because Galileo sort of sits in the middle. And so HR 2030 will help you with that. And we're going to be doing a lot more because now that it's built out and we're starting to talk to more and more vendors about it, we'll have vendor by vendor mappings in there and we'll continue to add more tools to gauge your maturity in various parts of the architecture. I think the reason I felt like I wanted to do this little shorter podcast is it's so exciting. What's going on, the potential.
[00:15:48] By the way, the other reason this is so exciting is the cost of AI is starting to plummet again. The Big two big frontier vendors want to go public, so they don't want to get crosswise with their financials. And the open source models are proliferating like mad.
[00:16:06] So some of you are going to have IT departments that are going to run your own AI and you're not going to have to pay giant token fees to these big frontier guys. And so you'll have a lot of the AI engines at commodity prices running on your own servers. So I think something like 70% of the tokens consumed around the world now today are in open source models, not in the two tier, two or three top vendors. So. And you can buy a harness like the Microsoft Copilot, which we love by the way. The Copilot is turning out to be one of our best platforms and you can run an open source model in it. So you don't have to pay the high prices of frontier model tokens if you don't want to. And that just means that the power of this stuff is going to be everywhere. We're going to have it in all of our experiences at work and all our experiences at home too. So it's really, really cool. Now, when you see the announcements that are coming over the next few weeks, a lot more of this will be clear. The second sort of part of it is the impact of HR on what I call business 2030. And I'm working on a big piece on that. We're not implementing AI in HR just for the sake of making HR a little more efficient. The real reason we're doing it is to implement dynamic enablement for growth. It's to allow the company to adapt and grow and reorganize and redesign itself more quick.
[00:17:31] Really, why HR 2030 exists, not to save money and reduce headcount. And we don't believe you will reduce headcount. You can if you decide that's what you want to do. But there'll be lots of new things to do that'll help your company scale. Okay, that's a little bit on that. We'll see you guys at Sav Connect. I'll see you guys at Workday Rising. I'll see you guys at Unleash. We're going to have a whole bunch of people at the HR Technology Conference in Vegas and look forward to seeing you all this fall. That's it for now.
