Episode Transcript
[00:00:00] Okay. Good morning, everybody. I just got back from the Middle east and I realized that over the last 90 days I've traveled almost 75,000 miles and talked to hundreds of companies, literally about their HR and AI strategies. And next Tuesday, I'm doing a large webinar to talk about trends for next year. It's not our predictions report that's still coming out later, but let me take a couple minutes and summarize what I've learned and you can join me next week if you'd like to listen to more. So there's been a lot of different studies being done on the adoption and use and maturity of AI in business, and I think they're all a little bit different. There was one a couple of months ago that showed that only 5% of projects had a positive ROI. But actually those are kind of long term projects and it's very hard to tell when they're going to have an ROI because they're big initiatives. I think the opposite, though, is true, which is that this has already gone mainstream. And every major corporation I talk to is using Gen AI in some fashion. Many of them have the Copilot or another standard tool. Many of them are starting to build internal chatbots. Almost all HR departments are using AI in recruiting, in various parts of recruiting. We're beginning to see massive growth in AI in learning and development. And that's our core. And we're spending a lot of time with companies going through that process.
[00:01:26] And I was in Japan, Singapore, Hong Kong, London, Paris, Abu Dhabi, Riyadh, Dubai, and every single city. I talked to dozens and dozens and dozens of companies that are doing interesting things. Now. A recent study just came out two weeks ago or a week ago from Wharton, another survey, and they found that 84% of respondents, and they're mostly leaders, are using AI every week, and 45% of them are using it every day. And I think it's fair to say with 800 million users per week on ChatGPT alone, to say nothing of the others, we're all using this on a regular basis. It is replacing our search engines, it is replacing our internal portals, and it will soon replace our learning tools, our employee experience tools, our recruiting tools, and many, many, many other things. And so if you read the article that I just published today, you can see that the major adoption that's taking place is for some of the simplest use cases of all, individual productivity. Now, you can't always measure the ROI of individual productivity, but we know it when we feel it and people are feeling it. It's easier to summarize meetings, it's easier to get through emails, it's easier to write documents, it's easier to write performance reviews, it's easier to find things, it's easier to create things, it's easier to create graphics or videos if you're into that. All over the place, individuals in all different walks of business, walks of life are finding productivity benefits to the Genai tools. And for those of you that are afraid of them, you're going to go the way of the dinosaur, to be honest, because this stuff is something you have to get used to. Now, when you look at the survey from Wharton and some of the other things that we discovered, that is not to say that this is an enterprise wide technology. By no means it is not. Very few companies have enterprise wide chatbots yet some do. I say it's about 10 or 15% at the most. Very few have good data management or data governance processes around these tools. Because the thing that's different about Genai from Microsoft Office or Excel or whatever the productivity tool you liked in the past is that it is very dependent on the quality of the data. So if you're using a tool that accesses internal information, it has to be accurate, up to date and timely or you're going to get the wrong answers, the wrong policies, the wrong systems and it's not going to work and you're not going to be happy. That means all of these portals, SharePoint sites, you know, piles and piles of versions and versions of policies and documents that are floating around inside companies have to be updated, governed, managed and kept current. There's nobody doing that in most big companies. I mean, it's done in a very haphazard way, even in small companies.
[00:04:16] So the AI gives us so much transparency inside of the company that we're going to be forced to clean up our act. Now there's much, much more to come, which you can read about in the article I just published because the single use applications that 85% of people are using are really very limited in their return on investment. These are typically not customer applications, they're not multifunctional workflows, they're not autonomy, they're not that intelligent. And as you can see in our four stage model, which I will go through on the webinar next week, and you can read about it in the article at stage two and stage three, the return on investment of AI is many, many times higher. I mean five to 10 times higher. And the analogy that I discussed in the article is a little Bit like a self driving car. The independent agent that you use for your emails is like a self driving car that only has a steering wheel assist. It has power steering. So you're the driver, you're deciding which way to steer and it's helping you steer, which is great. We've had power steering for a long time. Before power steering, it was actually really hard to drive a car. I remember cars that didn't have power steering. But that power steering doesn't tell you where to go. It doesn't deal with, you know, issues or bumps in the road. It doesn't steer out of the way if there's an obstacle. Maybe it has some beeps that go off to prevent you from hitting something. But what we want to do next is not only make us productive in our own driving, our own work, but make our work more integrated to the rest of the company in what we call multifunction agents. A multifunction agent is one that takes the work that you do and the work that somebody else does and somebody else does and integrates it together into a business process. And it turns out when we survey companies, there are very few of these so far. But this is where this is going in recruiting and training and development, in coaching, in the employee process. We're building multifunction agent capability into Galileo so that Galileo can be, and it already is, a digital HR business partner, a digital HR consultant, a digital HR learning consultant to give you training and education and coaching and tutoring, et cetera. And that's where this will have. These tools will have tens of times more economic value than improving your personal productivity. Now let's not, let's not undersell the personal productivity market. It made Microsoft a huge company. Companies like Atlassian and many, many other of these tools are fine for personal productivity, but the ROI on a personal productivity tool is a little bit hard to measure. And when you look at it, the cost of something like the Ms. Copilot, which is $30 per employee per month, it adds up very, very fast. And companies aren't going to spend that money if they don't see more economic value, which is why we still have very little direct revenue for these tools compared to the expense the industries of AI. All the various configurations of AI have invested roughly a trillion dollars in infrastructure this year. That's 10 months in power plants, Nvidia chips, data centers. It's becoming quite contentious where these data centers are going to go. When you're in the Middle east, they're very, very concerned about this because they don't have enough water. So they have particular limitations in Asia, they have other forms of limitations. So the expense or capital allocated to these massive systems is way ahead of the revenue. And so one of the questions, you know, that the world is asking is when are the big ROI projects coming so that companies can spend more and more and more money on this technology?
[00:07:59] Now, interestingly enough, in the Wharton's study, which is really just a survey, they concluded that 35 to 40% of the companies with large numbers of employees, I think 10,000 or more, are spending $10 million a year or more on AI. Now that might seem like a large number, but it's really not for a very large company with, with a large IT budget. The IT industry for software, not the services part, is around 6 or $700 billion a year.
[00:08:30] So you know, let's suppose 5% of that goes into AI. Maybe 10%, something like that. That will not even come close to paying back that trillion dollars of investment.
[00:08:41] Especially since the trillion dollars of investment has a depreciation of maybe two or three years. You gotta replace those chips every two, three, four years, maybe, maybe even more.
[00:08:51] Google thinks it's every two years, so there's gotta be more return coming. So in the article I talked about what's next? What's next is really two big things. Obviously. Number one is getting this stuff deployed as is. And by the way, specialized agents like Galileo are becoming very, very big tools because Galileo in particular is so well trained and so safe relative to HR that you can deploy it within your AI infrastructure and trust it without having to do all the data management yourself. And I think there's going to be a lot of those trusted providers, what we call vertical solutions that you're going to really, really love. But the next step is not only multifunctional AI, but getting these AIs to talk to each other. Now the MCP protocols and the A2A protocols that we're playing with are interesting, but they're not quite mature. And so, you know, a year from now these agents will talk to each other and you won't have to have such chaos between them. But it's certainly not there yet. And I think the risk we have to sort of consider in a larger company or even in a small company is, is if every employee goes out and buys the marketing AIs they want, the legal AIs they want the HR AIs they want, the recruiting AIs they want on and on and on, go through finance, administration, supply chain, et cetera. You can End up with a hundred AI agents in the company, 200. And you're going to wonder who's talking to who and what data's on which one and how do we keep these things under control. And this is not a governance process, this is an architecture. You need an AI architecture so that you don't buy a hundred or two hundred agents from a bunch of vendors that go out of business. But you have some integrated architecture and some process of deciding where to invest so the systems can talk to each other. This is all part of 2026 and 2027. Most companies are not dealing with this yet. A few are, but that's beginning to come up. And then the issue comes up of where's the data going to come from? Who's going to manage the data, how are we going to govern the data? We have a really fantastic story from IBM which is now in Galileo about how IBM does this. And you know, they have 6,000 HR policies and they're managed by different people who own each one. But it took them, you know, five, six, seven, eight years to get this all in place. You're going to have the same problem in every other part of your company. But you know, the upside of all of this new effort is you can shut down all these portals, you can shut down all these share point sites and internal applications like knowledge management and search and productivity and employee development and recruiting and internal mobility. All these things are going to be so much better and so much easier to do. And then as an employee, when you have a project or an activity or a task, you're going to be able to find an agent or maybe use the agent you've been given to do the thing that you want. I mean, we. One of our clients is a healthcare company that's actually had an agent for a long time. They worked with IBM before all this AI stuff started. And they use Watson. And what they've told us over the years, we've talked to them, we talk to them all the time, is that because everybody uses the agent inside of the company, there's no reason to build external, all sorts of portals and other systems anymore. More and more of the corporate applications are going away and getting plugged into this AI system. Now it's a proprietary platform, so there aren't a lot of interfaces and they have to hire IBM or work on IT integration all the time. But it's really helped them a lot because everybody can find what they need very, very quickly. Imagine how hard it is to do that today in a large company. Where 70% of the workers don't even have computers, it's virtually impossible. Now a couple of other things to mention and I'll be talking about this on the webinar. The vendor market is obviously in a huge frenzy. Every single vendor claims to be an AI vendor. You've seen the acquisitions by Workday, the acquisitions by SAP. Oracle's throwing around AI everywhere they can. UK announced a whole range of AI systems for frontline workers this week. Hibob is announcing AI. Hibob has now integrated Galileo into hibob. So anybody using Hibob can get Galileo directly linked and this goes on adp. I mean all of the big vendors are doing this. So those of you who are either running HR and it, you're going to be facing this three part question of do we build something, do we buy something from a specialized vendor or do we wait for our incumbents to build what we need? I'll tell you my gut feel on this, based on my experience doing this for so long, is don't wait. Because the big vendors will probably take time before their systems are fully competitive with the newer vendors. And then the intermediate sized vendors that you do business with, look for specialized solution providers who really know their domain. You know the reason that I think Workday acquired this company, Paradox is, is that Paradox wasn't a generic AI tool for recruiting. It was a very, very highly tuned, specialized system for high volume and frontline worker recruiting, which is a big market. They didn't try to do everything and that made them extremely good, just like we're not trying to do everything with Galileo. So I think you want to put your energies there and find some really best of breed solutions and then wait for the bigger vendors to build more or more interfaces. Because most of the bigger vendors won't build everything. They can't. They don't have enough resources so they're more likely to partner with or connect to the niche providers. But on the first category of building your own stuff, that's a big opportunity that we did not have in the past. I don't think very many of you ever went out and built your own payroll system or built your own CRMs. But with AI you can build a lot of your own productivity apps pretty doggone easily. And so consider the role of the builder part of this as you're looking for multifunctional applications now. One of the things I'm going to talk about in the webinar on Tuesday, and we'll be talking about a lot next year, is we've Been talking to so many companies about this. We were just meeting with several airlines over in the Middle east this week and a lot of manufacturers in Japan and others is that you really don't have to figure this all out yourself. We're going to help you, at least from our standpoint, with a blueprint on how these agents should be considered as clusters of functionality so you don't end up with a hundred things. And of course, you know, there will be vendors that will build more integrated, cross functional multifunctional agents over time and those are the ones that are the most likely to be threatening to the big ERP vendors in, in their internal applications. I don't think any of them are particularly big yet, but there's some very, very visionary companies out there, Eightfold Machi people, others that, that are going to be significant players in this market as you start looking around. Okay, last topic I want to talk about in the webinar came up a lot and I literally did have, I mean I, I'm not trying to, you know, overstate this. I think we had hundreds of conversations during this 90 day period is this idea of trust and fear. The consumer sentiment data that came out yesterday shows that in the United States, consumers sentiment is the lowest it's been in decades, the level of fear and employee engagement in the workforce is extremely high. Engagement is low, fear is high. And I think one of the reasons that Mamdani won the New York election is because people in the US at least are really concerned about the impact of technology on their careers and their lives and their standards of living. And you know, this isn't true in the Middle east, it isn't true in every country. But, but if you go around the world, you hear a lot of this. And many, many of the questions I got, particularly in Europe, were about trust. Is the AI going to ruin our careers? Is it going to make us dumber? Will we ever be able to hire candidates correctly when all the candidates are, you know, basically cheating on the tests and sending us fake resumes, et cetera. So a couple of things here. Number one, AI is not a human being. These AI tools. If you read the article I wrote about the BBC finding that 45% of ChatGPT searches are incorrect, AI doesn't know anything. It's a bunch of statistical vector calculus that is manipulating tokens. It doesn't even know what a word is. It simply intelligently recollecting and arranging tokens to respond to your requests. Now there's a lot of interesting math that will be added to AI as we get into more of what's called world models. But the large language models don't even really understand a number. They really only understand a token. So if you ask a large language model to do a math equation, it will write it out in English. It's the strangest thing because it doesn't really have mathematical capabilities. So if you think these AIs are going to be super intelligent and they're going to ruin your lives, I mean, they might ruin our our lives for a whole bunch of other reasons. They're not that intelligent. So first of all, I would just stop thinking about it that way and think about them as tools. I think about AI as a super duper duper spreadsheet, to be honest, which is something I learned about many, many years ago when I was at IBM. Okay, number two, will your jobs be eliminated? The answer is yes. Okay, I'm not trying to scare anybody, but I've talked to so many chros, it is not going to surprise me at all if a year from now the ratio of HR staff to employees will be 30% lower, maybe 50% lower. Now, does that mean 50% of the HR professionals in the world are going to be looking for jobs? No, it means you're going to be doing other things instead of doing gritty instructional design, running people analytics and tableau, pulling data out of the LMS to run reports, generating all sorts of documents to create an employee policy, or looking things up on the Internet to figure out what the latest legal regulations are, or reconciling payroll or et cetera, et cetera, et cetera. A lot of this routine stuff we're not going to need people to do. So if you get off sitting there doing routine work, you're going to have to look for something else to do. Now, I think most of you, And I think 99.9% of you probably would like to do higher value things. You'd like to be more involved with customers, customers with clients, with candidates, with managers. You'd like to do job design. You'd like to do more strategic performance consulting. You'd like to look at more interesting data problems. You're going to have the opportunity to do that, but it is going to eliminate or change a lot of jobs. Now, initially, it's really turning us all into superworkers because we're all giving, given the opportunity to use these tools to do the current job run better. But as soon as the agents do multiple functions, the job titles will change. So, you know, we may not need SDRs we may not need interview schedulers, sourcers. These specialized jobs, and by the way, there's 250 of them in HR alone, may not have to exist because you're going to be dealing with agents that do some of the specialized things and we'll be managing the agents. There'll be new jobs creating and created in data management, in analytics of these systems. We're going to get much more data about our employees. We'll know much more real time what's going on out there. We'll have much, much better data analyt. And so I would suggest that rather than worrying about AI as ruining your job, talk to your manager. And if you have a super manager, and this is what super managers are going to have to do, super managers are going to have to help each one of us think about our new roles, is look at new ways you can add value. Now if you're one of these people that's intimidated or afraid of AI, get over it, just try these things. If you want to get comfortable with the HR implementations, call us and we'll give you a copy. Or you can get yourself a copy of Galileo. Galileo is designed for you. It has four 400 prompts in it, each of which are so interesting. You could learn about AI just by running one of those prompts and just using it on your own company. And you'll, you'll be smarter about AI in an hour. So that's sort of my feeling on that now as opposed to, you know, these issues of AI making us more, less intelligent, taking over some of the tasks that we used to do. This gets back to these issues of cheating. In college or school, I had a number of conversations while I was out traveling with people that said, you know, my kids are going to be using AI to write their papers, so they're not going to know how to write, they're not going to know how to do research, they're not going to understand anything. Well, you know, you can't stop the technology from appearing. It's going to be here, we can't stop it.
[00:21:16] So we're going to have to rethink our problem solving domains. And my suggestion is that you think bigger about every single thing you do. If you no longer have to spend two days analyzing a spreadsheet to come up with a conclusion, and the AI comes up with a conclusion in 15 minutes for something that might have taken you two days of analysis, for example, first of all you got to check it. So you have to use your brain and your own analytic and skills and your own experience to figure out if it did the right thing. Because it may not have. And you're going to look pretty stupid if you use it and it was incorrect. Then you're going to have to decide what are we going to do about the analysis that was just created? I've done analysis now as an analyst for 30 plus years and I know what happens. You spend a lot of time manipulating the data, coming up with various findings, and then you suddenly sit back in your chair and you look at it and you have to go to a different place in your mind and you have to say to yourself, well, now that we see all this, what are we going to do about it? What are the implications of that? And that more complex thinking, which I call sort of business thinking, is going to be more and more a part of your job. If you automatically know why a candidate is the wrong fit, or why a person is underperforming, or exactly how somebody's skills are misplaced or misaligned with the work, your job isn't to do that analysis, it's to fix it, it's to solve it, it's to address the problem. So we're all going to move up a level to some degree in our jobs and be looking at the bigger picture. Even if you're a graphic artist or a marketing professional and you're building beautiful events or beaut marketing programs or images for the social media or whatever, I mean, now that you don't have to sit around with Photoshop and spend thousands of dollars building a simple graphic, you can look at the thing that it's created and really decide with your own human judgment. Is it the right message? Is it the right image? Is it the right colors? Is it the right brand that we want to promote? That's a human skill. The AI can't do that. So these are all situations that are going to get us out of the drudgery of what we do and move us up a level. And that's really, really empowering stuff. Now, if you're a frontline worker and you're dealing with customers, or you're in the hospital and you're a nurse, or, you know, whatever your job may be, you get to work at the top of your license. The concept top of license comes from the healthcare industry, where people who are trained nurses or trained physicians don't want to be sitting around moving furniture or writing stuff into forms. They want to be working with patients. If you're a truck driver, you don't want to spend a lot of Time filling out forms and dealing with safety issues in the truck you want to be driving, if you like driving. Ditto everybody else in the world. So what we're going to do in frontline work, as well as in white collar work, is we're going to be able to operate at the top of our license, which means you're going to have to decide what your license is. What is it you want to do with your career? What part of value creation do you want to be focused on in hr, in recruiting, in L and D, on whatever job you may have? You get to think about that. And there's going to be a thousand articles written about it, of course, and that's what AI is going to do. So I would ignore these issues of job displacement. Yes, companies aren't hiring right now in the US It's a good excuse, but I really think we're in the middle of an economic slowdown. I mean, I think there's a good reason why companies are not hiring, and that is they're worried about demand. And we have overbuilt the AI infrastructure to such degree that we're beginning to see a softening in demand. And that's just a normal business cycle situation that we're gonna have to work our way through. Okay, that's. That's it. I went for 25 minutes. Let me stop here. You guys think about this over the weekend. Come to the webinar on Tuesday. We'll probably make a replay available to those of you who can't come. And later in the year and early next year, I will produce and we'll be launching our predictions for the coming year with more specifics. It's all about the Super Worker, the Super Manager, and the Super Worker organization. This stuff is not slowing down. It's picking up speed. It's adding value. And it's a wonderful opportunity for you to enhance your career and your professional experience at work and the value you create. So stick with us, and we'll help you figure this all out. That's it for now. See you guys next week.
