Episode Transcript
[00:00:00] Hey everybody. Today I want to talk about enterprise AI adoption and what's going on in the labor market and some new information on what's really happening in corporate adoption of AI. Okay, so we are in the week of the introduction of the Venus release of Galileo, Galileo for managers, ChatGPT5, a massive number of new AI models released by Google. And it's very clear to me, and hopefully to you, that the speed of technology introductions is accelerating. So whatever you think AI is today, within six months to a year, it's going to be more. So we are sort of living in a world of technology adoption where we have this general purpose set of technologies that can do many, many, many things. Every everything from looking up information, analyzing information, generating information, creating graphics, creating objects, creating videos, actually understanding the spatial world around us, which is by the way, the new release from Google that you guys should look at and impacting jobs, careers, work, creative work, educational work, work in the analytics field, every sort of white collar and you know, human related job in some respects potentially could be automated, impacted, improved by this technology.
[00:01:29] So I stand by my philosophy that this is the super worker era and it's becoming more and more clear what these superworkers are. So what is the impact of this on the job market and what is the impact of this on companies? Number one, the job market. So I'm going to give you a link to some data from Goldman Sachs and there's a lot of people pontificating about this and I actually had a really good talk with the Department of Labor this week and we're going to have them on the podcast in the next couple of weeks. It appears to have a small impact on the job market. There are a lot of articles that have come out about entry level workers, you know, young people coming out of school having a hard time finding a job, the unemployment rate ticked up, the number of jobs created is slowed down, the time to look for a job is increasing. I've talked to quite a few of you I know that are looking for work and it's hard.
[00:02:20] And you know, you might think this is the AI effect, but if you listen to Goldman Sachs, they agree with the same thing that I've been saying, which is this is simply because companies are spending money on capital, not labor at the moment. They have decided, we have all decided that the more we invest in AI, the better for the long run. I mean that's what we've done here. We've spent the last two plus years investing in Galileo Learn and Galileo Agent, not really hiring any new people, but our business is growing significantly because of that and we can reach more of you at a lower cost. So everybody in the planet is going through that equation and they're slowing down hiring. So the slowdown of hiring of particularly younger people or entry level and even managers is not so much AI automating things yet, it's really budgetary. Now if you listen to the Goldman Sachs guys, they completely agree with this. I mean, Goldman Sachs believes that only 9% of companies have significantly implemented AI. The maturity model from ServiceNow that I'm gonna talk about in a couple minutes shows that about 15 to 20% of companies have implemented some level of AI. So we're in a very early days of this. And the package solutions for AI are yet to come. There's some, but not a lot. So mid market and small companies don't have the opportunities to buy off the shelf AI solutions yet. They will. So this effect is still early.
[00:03:48] And even though we have Dario Amadai running around, you know, screaming that 50% of jobs are going to go away, that's really not true. The Department of Labor agrees with me that it's not true because most of what's going on is we're automating tasks and many, many analytic activities at an individual level. And we haven't really built multifunctional, what we call level 3 or stage 3 AI yet. We in some cases have, but not as many as we might imagine. So the job market is going to slow down a little bit more and maybe we're going to have a recession. And one of the things the Goldman Sachs guy mentioned is that if we have a recession, that will accelerate the adoption of AI because companies will spend more money trying to figure out how to automate the stuff that they want to do to come out of the recession. So, so, so there is a slowdown in the job market, but we're not wiping out jobs. And there's an interesting podcast, I'm going to send you a bunch of links on that. The, the head of GitHub, which is the most sophisticated AI technology in any role of software engineers, basically says it doesn't really automate that much code. It writes code, but you got to go back and check it. And he believes that the software automation of code is just going to lead to more code and more AI engineering and other forms of software engineering on top of the code written by AI. Because actually if you think about the software stack we all use, even if we're software engineers, we use the code that was written before Us, the code in the operating system, the code in the database, the code in the networking layer, the code in the security layer, the code in the ui. I mean, we don't have to write any of that, right? But we're still spending, you know, a lot of money, 7% of the workforce, roughly, building AI tools and software tools on top of that. So stop thinking about the fact that all these jobs are going to go away. You do have to learn how to use these things. If you're a marketing manager and you don't know how to use AI tools, you're going to get stuck in the old school and nobody's going to want to hire you in a couple of years because all of these content things are being generated by AI. So, you know, some of us, particularly in marketing and creative and publishing and other things like that, we have to learn how to use this stuff. Okay, but it isn't a massive unemployment creator. And the Department of Labor agrees with me and you'll hear more about that soon. Okay, let's flip to the enterprise side. So we have a working group that meets every Friday, about 40 companies and we're learning a lot from all of them. And the thing that I think we're learning the most is the relatively early stage maturity we have in adopting these tools because they're not off the shelf solutions. There isn't like a guidebook or a user guide to AI that says follow step one, step two, step three. These are non deterministic intelligent learning systems. So we have to find use cases as quickly as we can where the maturity is high enough that we can use it quickly and get a return on investment. So I'm going to produce an article on this later over the weekend. You can read it. But the most interesting conversation I've had along these lines for a while is with Jackie Caney, who's the CHRO and AI Transformation Officer at ServiceNow. Now Jackie has a really interesting career experience. She worked at Anderson Consulting. She spent quite a bit of time as the Chro of Walmart. I met her there. She had a lot to do with their growth strategy in the 2000s and 2000s. And then she went to WPP and spent some time helping them organize their, you know, massive hundred thousand plus advertising business and then landed at ServiceNow. And ServiceNow, for those of you that don't know a lot about it, is the darling to me of the enterprise software market. It's a brilliant company run by brilliant people, growing at an extraordinary rate. A great stock, by the way. If you happen to be a stock buyer and very, very successful leadership team. And they understand the enterprise implementation of software and they understand how complex and difficult it is. And that's really their business is they built a layer of tools that sit between the transactional systems and the customers or the employees that run and build workflows and support systems and case management systems and knowledge management systems and other things including HR service delivery systems that don't get built in core ERP systems. And you know, you could argue that maybe ERPs become intelligent databases. I think Workday would debate that. But whether you want to buy that or not, ServiceNow is a massive opportunity. And they've been growing really quickly. And what Jackie and I were talking about was a whole bunch of things. But the thing that was most interesting to me was the work that they're doing on maturity of AI in general and what it implies about HR. So if you look at ServiceNow as a company, many, many things going on there, they have been able to grow in the high 20s at a multi billion dollar revenue run rate, which is very hard to do, but only growing their head count in the teens. So I'm not saying it's a 2 to 1, but it's definitely more of a 1.5 to 2 to 1 ratio of business growth divided by the number of employee growth. So they're proving to all of us that you can create a super worker company that scales up without hiring more and more people. And the whole idea of talent density and super worker, which I've talked about for many months now, is finding a way to grow your company at some rate, whatever that number is, depending on your business, and not hire people at the same rate. Not that it's bad to hire people, but there's a way to think about the organization to improve productivity on a continuous basis. That's what talent density is about and many, many ways to think about this. This is in some sense the core of the systemic HR model is that when we stitch together the HR practices that we have, we don't automatically hire people. Every time we have, we might develop somebody, we might move somebody, we might automate something. The old build buy bot or the 4R model that we call it is going into reality now at scale. So how do we get there from here? Well, there's two things that have to happen. There's maturity of AI in the company and there's maturity of HR too. Because if you're out there hiring people willy nilly through job wrecks that are flying in from all over the company.
[00:10:25] And the TA group is just, you know, fulfilling those as, as a fulfillment center. You're working against, honestly the AI strategy, which is to stop doing that and start automating things in some organized fashion. So that's why someone like Jackie as the AI transformation leader makes sense. Not that she does it alone, as I'll talk about in a minute, but we want the people, practices and the people and culture disciplines to link up with these AI initiatives to build growth, productivity and scale. We are sitting in the hot seat here in HR as a critical linchpin in the AI transformation. Now, what their research shows, and we're by the way, in the middle of launching a big study also on AI maturity that'll come out soon. And we're doing that in partnership with Workday. But basically what her research shows is that all of the things that matter in AI transformation, of all of the things that matter, the one that matters the most is the governance of the process. And I really believe that is true because, you know, there are essentially three ways of implementing AI in your company. Number one is you let the flowers bloom. You give everybody a tool, you issue a mandate, you train everybody a little bit and you say, go forth and automate and improve your productivity. That's a very difficult thing to do because individuals don't have control over their entire job function. They can do their own work better, but that's a relatively modest productivity increase, especially when we don't have sharing of information. So. So that let the flowers bloom approach, which is what's going on in most companies, goes to a point where you might get 10, 15, 20% productivity, but it's not clear that it scales from a business standpoint, because what we really want to have happen here is margin improvement and revenue growth, not just people having a little bit more time and doing more interesting things, you know, with their emails and so forth. So anyway, that's number one. And what happens in those experimentation stages is you bubble up use cases. And so the governance model that ServiceNow talks about, which we've talked about a lot as well, is, is these use cases bubble up and then there's a committee or a governance group that looks at the use cases and prioritizes them and decides where we're going to put more focused investment. And that's a really interesting group. And that leads to really the second approach, which is top down. The top down approach, when I talked to Tanuj from Standard Charter a couple of weeks ago, you guys heard her talk about it is, is sort of a little more like the CEO, cfo, operating leaders say we're going to automate this, we're going to automate that. We're a bank, we're going to automate the wealth management group, we're an insurance company, we're going to automate the claims process, we're a manufacturing company, we're going to automate the parts ordering process because we know there's a lot of cost there because we see a lot of opportunity for improvement and scale. That's just what we've decided. ServiceNow. Jackie and I were talking about the sales leadership function for example. They're a very fast growing company so they have a lot of salespeople and they have a lot of sales emphasis. Obviously Bill McDermott really spends a lot of time on that also. So, you know, are they, are their sales managers operating the most strategic way? Imagine you get, imagine, I mean I've been in sales my whole career. Imagine if we gave salespeople an intelligent agent that told them from the marketing department which of their clients are growing the fastest, which of their clients are asking for the most help, which of their clients are the most successful, which of their clients just got more money from the stock market. I don't know what all the inputs would be. We could make sales managers much, much more strategic than sitting around looking in Salesforce and trying to see if people made enough calls each week so those governance groups from the top down can take the bubbled up ideas and decide where to focus attention now. So then the question is, by the way, and the third model before I get into number the second one again is what I call the Elon Musk Doge model which is I don't really care where you guys do it, I'm cutting the people first and I'm going to strip you down to the bones and you're going to use first principles and you're going to build back up your organization in a way that'll make it more productive. One way to improve productivity is to constrain resources. That's what works in startups. I've been a startup guy for, I don't know, almost 30 years now and it's nice to have constrained resources because you have to make difficult decisions. You can't do everything and you actually redeploy people very quickly and it's actually really fun and it works great. You know, it forces you to think about skills and capabilities and alignment and culture just like everywhere else. But you can't just hire, hire, hire every time you need Somebody. So that's the third model. Some of you are in companies that are, I'm sure going through the third model too. So let's go back to the second model. Let's assume you're using the bubble up approach and you're getting ideas and you know, you're getting all sorts of really clear use cases. Well, this governance group or this governance team needs to deal with four or five issues. The first is the business benefit of this application and that's probably the CFO and the business line of business leader in this governance council saying, you know, of all the things I'm thinking about in my group, this is number one. This is the one that can have the most impact. The second is the data governance. Do we have good data? Can we get good data? Can we protect the data? Can we consolidate the data? If the vision of this group is to do something really extraordinary, but the data is all over the company and not in one place. It's not going to be a simple project. And even if it's worth the investment, and it may be, by the way, worth, you know, getting all the data together for this, someone's going to have to lead that and someone's going to have to own it. This, by the way, the reason Galileo has been so successful is we happen to have a very good data system. We have really good consistency in our research and our tagging and our data quality. We were just fortunate of doing that because I worked for a database company for a long time and just have a data head. You know, that's not typically true in a big company. The third is bias and ethics.
[00:16:38] Someone in this governance group needs to think about what is the risk of this new system going sideways? What is the risk if it's making the wrong decisions? What is the risk if it becomes biased? Is our data biased and what do we need to do to prevent that? Most of you know that there's a lawsuit that could be huge or could be nothing against Workday for bias in hiring based on their technology. You know, I don't think Workday has anything to do with what is going on. But you know, this may end up turning into a big issue anyway. But you don't want to run into one of those deals. So if this project is a big project, somebody needs to look at that. The fourth is the human capital issues. If we get this new system to work and we build the prompts and we find the right technologies and we get the right data and we have the right business context and the workflows defined? How are we going to set up the team? What is the role of the humans? Who's going to do the training of the AI? Who's going to monitor it? What are the skills that are going to be needed that we don't have in this group? Who's going to get an upgraded job? Are we going to have to hire people? Are there new roles here? Yes, there are, and therefore are. Do we develop people for those roles? Do we hire those roles? And that's the HR side of it. And then there's the IT infrastructure side, which is, do we have the tech to do this? Which of the agents is the best? By the way, each of the agents is different. And so most of the companies I talk to that are relatively mature have found that no one agent does everything well today, although that's kind of what ChatGPT5 is trying to deal with. So, you know, maybe there's a specialized agent we don't have yet, or another technology we're missing and, you know, which of the vendors we trust and all that stuff. So this governance committee that takes the ideas or builds the ideas has to do all those things. Well, what the ServiceNow studies have found, and this is really interesting to me, is that if you look at the overall level of maturity of AI in organizations, and it's not very high, you know, Goldman Sachs thinks it's fewer, less than 9%, and the ServiceNow study shows that it's around, I don't know, 12, 13%, something like that. I think it's very low because there's so much yet to come. It's probably lower than that because these technologies are going to get better and better and better so fast. But anyway, the most important correlating factor in maturity is governance is this process of finding opportunities, exploiting the opportunities, leveraging and developing the AI solutions and then rolling them out in a form that the user community and the enterprise and the staff and the employees and the customers can adopt them. So this is not the type of transformation we did when we bought SAP or Workday or SuccessFactors or wherever it may be or an LMS. And we just sort of put together a project plan by Accenture or Deloitte, and they just implemented it. This is not that kind of work. These are creative tools that are changing all the time. And whatever application you decide to go after, you got to be also aware of the fact that the technology is going to advance very, very quickly. So get ready for a project that's going to have a life of its own just Galileo alone. I mean, every quarter we get new technology into Galileo, I'm sure ChatGPT5 will be in Galileo within a few months. And we go back and we look at it, and we got to look at all the prompts we've built and all the stuff we built and make sure it all works well, usually it works better. So, anyway, those are some things to think about. Now, for most of you in hr, this is all exciting and a little bit intimidating. I would just offer you in terms of working with us. Galileo and Galileo Learn are now bundled together. And so if you go to Gegalileo AI, you'll see how this works. Our AI stuff is really a great education tool and enabling tool for you, so I really want to encourage you to get your hands on it. We can license it to you as a team, as a group, as a company, and it's the vehicle that we're using to reach you and to do the advisory and research work that we do. By the way, for those of you that don't know what we do or you're confused about what we're offering. Just to make it clear, I just put together a video yesterday to explain our business model and why and how AI is transforming us. And I think that'll be hopefully useful for those of you that want to work more with us, but also to just think through that. AI doesn't replace what you do. It turns you into a super company, super worker, super team. And that's certainly the way we see it. Okay, so I'll put a bunch of links into this podcast so you guys can read some of this stuff and listen to some of these other things, and we'll talk more next week. Bye for now.
