Episode Transcript
[00:00:00] Hey everybody, it's Friday, I just got back from London and I'm going to give you a big update on two things this week. First, the state of AI and the threats it now poses to hr. And I don't mean that in a negative way, but I'm very, very now aware of what's going on after meeting with a whole bunch of companies over there.
[00:00:24] And then I'm going to touch on the revolution of learning and development which you've heard from me before. But this week I had a pretty big session with, I don't know, 500 or 600 people at the Learning and Development Technology Conference in the uk, which is, I think the largest conference of that type and people really, really ate it up. So I'm going to give you a little update on that and then much more to come around the middle of May.
[00:00:52] So let's talk about AI first.
[00:00:55] So I've met with at least 30 companies in the last 60 days, maybe more. And this week I met with at least 10 in various settings face to face. And there's a pretty unanimous issue going on that you are being asked in HR to evaluate and implement AI to both improve the productivity of HR and do things faster and cheaper, but also to help the company implement AI across the various implementations in the company. And as much as it may sound a little bit inflammatory to say this, I think this threatens the HR profession as we know it. And I'm very aware of the value of what we do. Believe me, I'm a, you know, a huge part of this. But I think what is happening, and this is going to probably get a little, maybe a lot worse before it gets better, is a lot of pressure from above, CEO, CFO and others to not do things as usual to improve productivity. And by the word, you know what the word, you know, the veiled threat of productivity means layoffs or headcount, headcount, implement agents, implement the co pilot, reduce expenses. And by the way, the IT people are bulking up to do this. I read a study last week that said that IT budgets are up 60%, 62% to buy tools for AI. And you know, a lot of the companies I talked to had the Microsoft copilot, they weren't doing much with it, but they had it.
[00:02:33] Several of the companies I talked to have built custom AI solutions which by the way is turning out to be a little bit of a better strategy because then you have complete control over what you're going to do with it. And then there's specialized systems like Galileo that we're very aware of. And Galileo is exploding with growth. I mean, I don't remember how many people we have using it now, but it's four or five hundred companies and we're beginning to get all sorts of feedback on what, what people are doing with it.
[00:02:56] So you're in a sense under pressure, whether you like it or not, to find and implement AI solutions as quickly as you can. And of course there's so many opportunities to do it in hr. It's low hanging fruit everywhere because we've got all sorts of paperwork things and use cases and things we can do. And so the question I keep getting is where do we start? What should we do? What vendors are the best ones to work with? So let me give you a couple of insights on that based on what's been happening this last week. First of all, all the vendors have AI stuff, every single one. You know, some of them are early stage, some of them are a little more mature, some of them are used for sourcing and recruiting things like talent intelligence systems, agents for recruiting, agents for onboarding, agents for coaching, agents for leaders, agents to help you with call center inquiries and self service for employees on various tactical stuff, agents for performance management. I think I mentioned that last week, is turning out to pretty big deal because that's a fairly easy problem to solve and that makes things a lot better.
[00:04:03] And then of course the internal agents inside the HCM systems to make the HCM system work better and manage various talent practices. And you know, my 100% certainty is that over time we're going to have AI agents that do all of this and they're all going to either talk to each other or they're going to come from one vendor. Now this isn't happening yet because the vendors are working on one piece at a time. But if you really think about why we're here in hr, why we even exist, we exist to manage all these human capital practices and processes in companies, each of which are very related to each other and each of which have data that can be used between them. Right. I mean when I hire somebody I learn all sorts of things about their background and their skills. And so I kind of know how to onboard them. And once I onboard them and they start working, I know where they fit relative to other people I onboarded in the same job and so I know what I need to do to coach them. And then as they progress, I know what they are good at and not good at and what new job they should get and what new role they should get. And on and on and on. So there's, you know, a huge opportunity here to completely change this profession. And so I'm working on a big article on this. So I won't go through the whole thing right now. But there is a threat and it's real that we're going to wake up two or three years from now and you're going to look around and it's not going to be the same HR department at all. It's going to be very different. And I'm pretty sure I know roughly what that's going to look like. I can't tell you exactly, but unfortunately HR is not the number one priority for AI in most companies. In fact, it probably shouldn't be. There's much more important things to do. If you're a bank, you probably want to have an AI systems for wealth management for your largest wealth management customers. You probably want to have an AI system for account opening and account management. You probably want to have an AI system for risk management and risk assessment. You probably want to have an AI system for determining if somebody's qualified for a new loan and then issuing that loan and doing all the paperwork around that and then verifying, you know, all the things you have to verify to close that loan. I mean, all of these things we do, and I'm just talking about a bank, all of these things we do in companies that are transactional in nature can be automated and integrated in just like we can the various integrations in hr. So how do you decide when you get this mandate, where to start and what's going on is a whole bunch of things at the same time. Some people are just playing around with gen AI tools one at a time. At level one in our model, which you guys have all seen, some of them are looking for level two tools that are more automated, more like Galileo. Some of them are looking for level three tools which don't exist yet. There just aren't very many that can go cross functional yet. Although they're coming, we talked to, I mean Paradox is really good at cross functional recruiting. Maki people who I had dinner with this week is an incredible tool for end to end recruiting and they're going much further than this. And then, you know, and then in the L and D space where I spent, you know, a day and a half, there's tools now available that can really revolutionize everything. Reduce the need for instructional design, reduce the lead for an LMS or eliminate it almost and reduce the need for content development and content management and content analytics and all those other things we do in in L and D. So you as an HR person or you as an HR team or a leader have just dozens of things you can do. And one of the things I'm going to talk about at our conference is what you know, how do you manage this list of projects? But I think there's some experimentation that has to happen so everybody knows how these things work. You have to do a good sober look at the vendor market and get our help because a lot of this stuff is immature. And then you're going to have to prioritize. Where are the opportunities to add the greatest amount of value to the rest of the company? What are the problems that we want to solve? Getting back to this long discussion of falling in love with the problem, don't fall in love with the tool, fall in love with the problem. Because the tool might be great, but if it doesn't solve a problem, you're going to have a hard time getting any money to work on it when there's, you know, 50 other AI projects going on at the company at the same time. And we are starting a project, by the way, right now, Kathy and I are going to lead this to look at governance of AI, which means how do these AI things work with each other? Because they're never, there's no standards for how they interoperate yet. So even if you find some stuff you love, they're going to be little islands of automation and you're not going to necessarily be able to connect the data together.
[00:08:45] And by the way, that doesn't mean that if you buy workday, everything's going to work either because their different AI solutions are embedded into the functional areas that they are built in in workday. So anyway, so that's a massive sort of priority for companies right now. And then in the middle of all that is something that's even bigger, which is that every employee very, very soon is going to have their own large language model. You know, if you look at what Microsoft's doing with the Copilot, you look at what ChatGPT is doing, you look at Galileo, you look at the ServiceNow now assist, you look at Joule, you look at the workday assistant, those are essentially tailor made agent or chat systems designed from the vendor in which they were developed to do the things that that vendor does. Well, nobody's going to want 20 of those. And everybody knows this already, we're going to want one. And if I'm working in a store or in a plant or driving a truck or out on the road talking to clients. I don't want to have five chatbots, I don't have time. I want one and I want that one to be mine. I want it to know my calendar, I want it to know my emails, I want it to send me messages, I want to it questions when I can't find something. You know, when you're traveling, I want to know what time zone I'm in and what meeting I'm supposed to be at. And by the way, how do I, you know, where's the car that's supposed to be meeting me? Right? I just want my own little assistant that does all of that stuff tailor made for me. And by the way, if I'm a barista in a store, I want one. And if I'm a truck driver, if I want one and if I'm for salesman, I want one. And these things are going to be in your phone, so you're not even going to know it is an lmm. It's just going to be there. The new, the new Sana app that we've connected up to Galileo, which you can download from the Apple App Store, basically turns Galileo into a mobile app. And you can talk to it and it talks back to you. And that means that the architecture of all of these systems isn't like it used to be with Workday, Oracle, SAP, where everybody had to log into a big server and use the cloud. We're all going to be interacting with all of these systems through our personal LLM. Now that's not happened quite yet, but it's, you can just see it coming. And so the architecture for what we're working on with Galileo is a two tiered architecture where Galileo is the front end to a whole bunch of other things that are going on behind the scenes to make you a better HR professional or a better leader. And that's the way you have to think about your company. So in some sense we're playing into Microsoft's hands or whoever you decide to use because they have software on 70% of the desktops in the world. At least in the business side, Apple doesn't have anything like this. But you know, those of us that have iPhones are kind of wishing that the LLM was in the iPhone as opposed to having to download an app and click around from one to another inside the phone. So that's the world we're getting to and we're going to get there very fast. And the reason I think this is going to happen fast is the pressure is on, you know, the tariffs are out there. Every company's worried about the economy. I don't know if we're going to have a recession or not, but there's a lot of evidence that we are. And virtually every company I talk to is doing some form of work or job redesign project to figure out how to do more with less. So let me talk about that. In the second sort of discussion about why we're implementing AI is the issue of how do we redesign the work to use the AI. Now, the reason we have unproductive companies goes back to the fundamentals of how we run our companies. The reason we have companies at all is that a group of people working on a business of some kind is more productive and more efficient than a bunch of single people doing it. If you had a bunch of single people trying to build shirts or design and manufacture shirts, they might be nice shirts, but you're not going to get very many of them per month. But if we put them into a supply chain and we have a designer and a manufacturing person and a repair person and a person who worries about logistics and supply and so forth, we can build a company that makes shirts, and we can create 10 times as many shirts per dollar, per hour, per level of input. And we can probably iterate faster on new designs and be more creative too, because each one person can only do one thing at a time. That's the story of why we have companies at all. Now in the industrial age, before we had all these thought workers and information workers, we designed these companies as machines and we had inputs and outputs and we had workflows and factories. And the jobs, the job architectures that we now still use were designed to create this machine like industrial process for building, marketing, selling to, supporting, distributing this thing, whatever it is that we're making or this service that we're delivering. And so we had very rigid jobs. But of course, what happened, you know, as we got more into services and intellectual property and recurring revenue things, and rapid disruption of various businesses, industries combining with each other and so forth. You guys heard me talk about this, you know, for a couple years here, the industrial job model started to decay. And we ended up with a situation where there were a lot of jobs created that were not necessarily industrial designed.
[00:14:17] So the reason we now have productivity problems in a lot of companies, and the reason I say this is because I've had a chance to talk to them, is when you listen to the podcast I just recorded with WPP, you'll hear the story. WPP, the world's largest advertising agency has 100,000 employees and about 60,000 jobs, which means every employee almost has their own job. It sounds insane when you say it, but many, many companies are like this. We have another client that showed us their job architecture. And every single functional group in the company has a series of analysts, a bunch of project managers, a bunch of program managers, and other staff overhead that's been replicated into all these different groups because each group decided on their own when they were given more headcount that they wanted these overhead jobs to make their work easier. Nobody sat down and said, let's build some plumbing so everybody can use the same plumbing. Everybody said, I'm going to build my own well, my own hot water heater, my own piping, my own pumps. And by the way, you know, my pump's like this. And you're going to buy that kind of pump, but I don't want your pump because I don't have a piece of pipe from your side to my side. So I'm just going to build my own pump, right? That's kind of why we built all these companies. And we never really taught managers or leaders. I certainly never learned how to do this, how to think about the organization as a flow. We think about it as a bunch of human beings. And I interviewed nicolemeroux from IBM a few years ago when we were first doing our org design research. And I asked her, you guys are a very, very sophisticated global company. How do you do org design? You must have a methodology for this. And she goes, no. At that time, she said, we don't look for the executives that are the most successful. We put them into the jobs where they're the most needed, and we let them decide how to manage the organization. So anyway, so even the best companies in the world have lots and lots of people doing lots and lots of things that are not designed in the most productive way. Where we're sharing resources, sharing skills, scaling at a rate where we don't have to hire more people to scale. And actually in the opposite happens where because we have a lot of duplicate roles, we actually become less efficient the more people we hire. Because the more people we hire, the more people you have to talk to, the longer the staff meetings and the more problems we have with who's accountable for what. Because there's a whole bunch of people that are sort of doing similar jobs, right?
[00:16:48] So this AI push that we're getting from the C level is forcing all of us to take these works and jobs and roles and skills and things we've got and clean them up in some manner.
[00:17:02] Now, I like to talk about Elon Musk because he's such an interesting guy. So his philosophy about productivity is we strip the whole company down to the bare minimum where it barely works at all, and then we incrementally add very tiny amounts of resource to make it scale. And, you know, he does this because he's an entrepreneur and his companies are not that old. They're relatively new. So, you know, the reason that SpaceX was able to get a rocket into space at, you know, like 1/5 or 1/10 the cost and the weight of NASA is he kept saying, they blew up a lot of rockets. They had a lot of them that failed. He kept saying, well, get rid of that, get rid of this, get rid of this, get rid of this, get rid of this. Find a cheaper way to do this, find a cheaper way to do this. Until the rocket blew up. And he said, oh, wait a minute, add that back in. And it talks. He talks about that in the book that Walter Isaacson wrote. But we don't have the opportunity to do that. If you're a bank or an insurance company or, you know, a big manufacturer, you can't sort of break everything to see what the bare minimum requirements are to do it. You got to engineer it, you know, carefully. Now you have to decide where can we put in plumbing? By the way, I'm using the word plumbing because Tanuj, the chro of Standard Charter, I spent quite a bit of time with this week, loves to talk about plumbing because she's, she's actually the chief strategy officer, is that the better the plumbing, the faster the company can grow. And so a lot of the work that we have to learn how to do, by the way, most of us have never done it or not very much is what is the plumbing. We need to reduce the number of people and bring in these AI agents so the AI agents can be used to streamline various processes. Now, three, five years from now, the agents will be mature enough that you'll be able to buy the agent and decide, using the agent, what the workflow should look like. But the agents are still new, so we're building a lot of this in real time. So this was every single meeting we talked about this stuff. I mean it all from top to bottom. And what I'm going to do at Irresistible, I was thinking about this on the way home is I'm going to work on a speech and I'm going to talk about how to build a super worker organization, not how to be a super worker as an individual. But how do you do this? And I'm going to spend a good 45 minutes on this based on what we've learned and give you guys some things to think about. So stay tuned for that. Okay, so that's a little bit of what that's all about. And that's a quick update this week. Now, L and D. The L and D Technology conference in the UK is the largest of its type. There's about, I don't know, 10,000 people there, 8 to 10,000 people there, all sorts of training professionals, training managers, chief learning officers, training administrators, training tech people, content developers, all sorts of people. And the learning and development industry, which I love and I spent so much time doing this, is filled with really smart, really academically interested, technology savvy, creative, hardworking, passionate people. And as you're going to see in the announcement and the new big research report we're publishing in May, this wonderful, important $360 billion industry has really, really stagnated. And I won't give you the whole narrative in this podcast, but we have been stuck in what I call the publishing paradigm. The publishing paradigm is pretty obvious when you think about it, but is the way we've run L and D all the years that I've been in, it is we have responsibility for training people, certain group of people on certain thing.
[00:20:48] We go through some form of needs analysis to figure out what they need to know to do their job. Well, that in and of itself is a science. We get constant input on that. We do surveys, we interview people, we do on the job activities and so forth. And then we sit down and we build a training experience or intervention or program or simulation or whatever you want to call it, to teach people how to do that thing so that they can do it better and the company can grow and they can grow as an individual.
[00:21:23] Now that is a very complex process. Because teaching somebody how to do something doesn't mean just teach them how to do it. You've got to teach them about it so they know why they're doing what they're doing. And because you can't teach them every possible use case of how they're going to what's going to happen. You have to give them enough knowledge and experience that they can deal with all sorts of things that come along the way. So if I'm teaching you how to build a coffee at Starbucks, by the way, you know, the ex clo Starbucks I hung around with a lot there, he said, basically there's A hundred thousand permutations of coffee at Starbucks. So you can't teach people how to build a hundred thousand things. You can teach them how to create things in different forms, but they have to learn how to build these themselves. So you've got to give people an opportunity to practice, you've got to give them a coach, you've got to give them feedback, you have to give them the opportunity to ask a bunch of questions. Because everybody comes from a different background when they go into a new job.
[00:22:25] And then you have to give them the freedom to do the job the way that they want to do it that's best for them, yet still achieve the quite same outcomes. So this thing which we call training is actually very sophisticated and very complex. But unfortunately, because of the learning and development architectures that we've had in the tools and the old standards, we treated it as a publishing problem. We developed a course or a video or a wallet card or a checklist or an electronic device of some kind that taught people what they needed to know. And maybe it was a performance support tool that actually monitored what they were doing and helped them finish this, this thing they're trying to do without them having to learn everything. And then we published it, in other words, launched it and then we deployed it and then we encouraged people to use it or consume it, or take it if it's a course. And we crossed our fingers and said, wow, that was a lot of work. Let's hope this solves the problem. And we would survey people and go out and do our ROI analysis. And I wrote a book on that. And that is a batch process. That is a process of needs analysis, design and architecture, content development, or program development, delivery, which may be electronic or maybe in face to face or maybe otherwise getting the people who do the delivery to know how to do the delivery, launching and, and attracting people to consume. This thing you just built and then manage, measuring and managing whether they learned anything and whether that actually resulted in outcomes that you wanted in the beginning. That is a long batch process. I have a chart on that. I'll show you guys the slides. When you come to irresistible, it takes a long time. And during the process you're doing that, you run into a whole bunch of things. We need it in multiple languages. We need a different version of it for this department because they don't do it the same way as that department or this country or this city or this division does it slightly differently. So we build the, we build the core thing and then we give it to a local Training department and they tweak it or they add cultural versions of it to their look from their local group. And it's a big, big, complicated operation. And the reason it's a $360 billion market is that in order to do all of those things, we need a lot of tools. We need a lot of content generation stuff. We need, you know, pretty fancy lms to track it all. We need measurement tools, we need assessment tools, we need some video tools, we probably need some character generation tools, some simulation tools. So, you know, companies end up buying a lot, a lot of things, all each of which are cool in and of themselves. And they've got this publishing paradigm. And then there's a whole bunch of people in the middle that are publishing stuff, editing stuff, creating stuff, trying to reversion it. And you end up with L and D and a lot of people in L and D and a relatively slow process.
[00:25:24] Well, apply AI to that and 70 to 80% of that changes. Because with an AI system, the needs analysis is almost automatic. Because if the person has access to a chatbot or their own local LLM, as I mentioned earlier, they're going to just ask it a question. How do I do this? How do I build this kind of a drink? Or how do I deal with this problem employee? Or how do I deal with this customer that won't buy this thing that I just showed him? Or how do I make this demo work better? Or whatever. And that input is now instantly available to you as an L and D person to quickly learn what the issues really are that are holding people up as a group and as an individual. And now instead of going back to the batch publishing process and saying, ooh, let's change the course around, let's build a new version, let's build a new chapter, let's build a version for this type of person versus that type of person and try to personalize this. You let the AI build the content and let the AI answer the question in different forms. And that is what AI can do. But we didn't design L and D to work that way. We designed L and D for this daisy chain, manual, publish, produce, market, measure, iterate process.
[00:26:43] So a lot of the jobs in L and D are going to either vanish or get completely changed. And that's what I was talking about for an hour or two with everybody in Europe. And I'll tell you, you know, what happened is people were fascinated by this because they never thought about it. I mean, I think they see AI tools coming into the market and they see AI vendors and they talk to the AI vendors. But once you get your head wrapped around this paradigm change, you realize that this really is a revolution. This is not a step change, better way of doing things the way we've already done them. This is a different way of doing things, a very, very different way, exceedingly different way. And so that was the other part of my week, is talking to a whole bunch of people about that. And what comes out of it is when you sit down with somebody who's working in a given company, in a given domain of their business, certain audience that they're trying to train, you start to get into, well, what do they do and what systems do they use. And maybe we can plug this AI thing right into their flow of work and it can start to give them tips and suggestions and feedback while they're doing the work, in addition to training them before and after. And we can give them access to coaches and mentors and AI tutors and all sorts of cool stuff like that. This is really different. And when you read our research report in May and you see what we're working on next, you'll see how big this is. So that's kind of what happened to me this week. You know, generally speaking, when I go on these kinds of trips, it is incredibly exciting for me because I get a chance to see how real these problems are. And I would say in both cases, in the case of AI and automation and tools, and in the case of L and D and reorganization and revolution of L and D, we have basically the same problem, which is that there's the fear factor of wondering what your job is going to be once this world changes. And then there's the understanding and learning factor of how do I become a part of this? And I would say that among the steps I'm going to explain on becoming a super worker organization, the number one is you have to experiment and get to know how these things work. Every good implementation of AI I've run into, and I've run into a lot, the customer, the company, the people did a lot of playing around. They, you know, decided what technologies and tools to use. And then they learned that they were really, really good at some of the things they thought they were going to be really good at, and they weren't very good at some other things. And they had to work with the vendor very closely. And that's just where we are. The good of all of this is that as we come up with the reinvented model of HR&L and D and Businesses in general, we're going to come up with a lean and mean operation. And I like that phrase lean and mean because lean and mean is not a negative idea. It's a positive idea. The idea is if we can do this better, faster, a higher quality for our customers, our stakeholders, our shareholders, we get more value out of it. We're more, having more fun, we can probably make more money because we don't have as much overhead in the company. We can probably make more money as individuals too. And we can iterate and improve and innovate faster.
[00:30:02] So you know, you can look at this as an AI revolution. You can look at this as agents, you can look at this as the magic of LLMs and reasoning models, or you can look at it as the, the business re engineering opportunity of our lifetime. And what I said to everybody in the L and D conference is I said, you guys mark your watches. Write yourself a Note that in 2025 you entered a world of reinvention that was maybe the most significant reinvention of your career, your company, your role, your operations than you have ever been through in maybe your entire career. And this is the beginning of it. So that is pretty much my trip report on what happened when I was in Europe. So the next couple weeks are pretty big. Next week we're at the Unleash conference in Vegas. We have 10 or 15 Galileo workshops. If you're coming to Vegas, there's still a few seats left. You're going to get hands on experience with Galileo, another thing that will be good for you and learning how to do other, other forms of AI. Then we're doing a few more things. And then the week of May 19th is our conference irresistible. Irresistible is almost full. There's a few seats left. Bring your HR team.
[00:31:18] Come as a leader ready to have a good time and learn a lot. We're going to be launching some things there where you have some amazing, incredible, very, very talented chros coming, some external speakers you're going to love, some very special events. And it's a wonderful, wonderful experience to go to usc, one of the most beautiful campuses in the world and a nice warm experience in Southern California and enjoy being with some of the greatest people that you will love spending time with, including us, of course. So that's kind of the latest this week. I hope this was at least a little bit of inspiration and maybe a little bit of learning for you this week. And I'll give you some much more specifics next week about what's happening next. Talk to you later. Have a good weekend.
