Episode Transcript
[00:00:00] Today I want to talk about the economics of AI and the job market because we had a very interesting unemployment report or jobs report that showed 100 and some odd thousand jobs were created, which was considered to be bullish. But if you look at what those jobs were, they were mostly healthcare, a third of them were healthcare and the restaurant, transportation and distribution. No jobs were created, in fact negative in information services and other parts of the economy.
[00:00:28] So it leads me to think about some other things that happened this week. The big announcements by ServiceNow and the discussions of the SAS apocalypse.
[00:00:38] And I want to sort of spend some time this morning just talking about the economics of all this. So we are, we are definitely poised on what I consider to be a trillion dollar re engineering or redesign of businesses in many, many ways thanks to AI. And there'll be other markets for AI besides corporate. And that trillion dollars is going to be spent on redesign and technology development of entirely new business processes and systems that integrate many of the siloed functions in our companies into more customer and employee based experiences. So instead of you having to call a customer service agent when your flight maybe didn't take off and the customer service agent trying to figure out what happened and the people at the airport putting the wrong into the system, like what happened to me this week when I flew to Florida and all the information from United Airlines was incorrect all the way to the point of reaching San Francisco and what happened after that, all of that would be flowing to me or the customer in an integrated form because the AI integrated systems would show you the big picture of what's going on with your product or your service or your flight or whatever it may be. And you know, companies like Amazon do this. Apple does a bit of where you do get a complete picture of your relationship with this customer, with this company and it makes you want to buy more stuff. And that is really hard to do today because these older ERP systems and ORG structures we have in companies weren't designed that way. They were designed more for scale and manufacturing. So this is going to be a lot of money, it's going to be a lot of time. And in HR we have lots and lots of these siloed functions. In some ways, as I was describing yesterday to a bunch of chrosis, HR 2030 is about one simple idea. Dynamic enablement of every employee and worker in your company.
[00:02:36] Dynamically real time, giving them what they need to do more, be more productive, skill themselves, grow, be happy, contribute, et cetera, without the HR department behind the scenes trying to design things in a more batch process to, to create that dynamic experience. And it's going to take a decade to see this play out, because there's going to be a lot of innovations in the tech and in the business thinking around it, but it's absolutely going to happen. So now the question is, how is it all going to get paid for? And we already know that probably, I don't know, maybe a trillion dollars has been spent on data centers and chips and infrastructure and interconnect and power plants. And now we're going to put power plants into space. Because these AI systems take a lot of computing. They're probabilistic. They don't just transact what you, they don't do what you tell them to do. They figure out what to do. So they have to do a lot of thinking and consume a lot of resources to do that. So economically, the money that we, we spend on human thinking in our companies, the interconnectedness of our human business relationships, is going to be spent on electricity and digital thinking by the AI. And, you know, the AI is better than what we do in some respects and much worse in other respects. But that being, you know, let's put that aside, that's a big shift in resources. Now, if you think about it from the standpoint of a CEO or a cfo, they look at the, you know, P and L on Wall street and they say, okay, you're growing at this rate, you're making this much money, and here's your expenses. If you can tweak the expenses down and the revenue up, we're going to, you know, up your stock and you guys are going to make more money and be more successful. So what the business community wants to do is offset labor cost with this computing cost. But we don't really know what the computing cost is yet because the computing cost providers and the software and AI providers haven't really figured out how to make money and how much they want to charge. If we find out, which we will soon find out, that Anthropic and OpenAI are losing a huge amount of money on every one of the $20 a month subscriptions they sell us, then the jury's out whether we're not going to be able to afford this as much as we thought because they will raise the price of what we have to pay and the economic balance, or teeter totter, will change because it won't make as much sense to replace so much labor with technology if the technology is very expensive. And the big example of this was this massive amount of information that came out from ServiceNow this week. ServiceNow has these beautiful charts where they show that you triple your spending on ServiceNow, you'll still save a huge amount of money by reducing 80% of your labor. Well, I mean that's nice in a chart and a sort of beautiful graph, but that's not actually how it's going to happen. I really don't believe any of you are going to get rid of 80% of your people around AI anytime soon. Maybe at some point in time that will happen, but those people will be doing other things.
[00:05:46] So there's a lot of uncertainty about how we're going to pay for all this. Now in the middle of all that, we have this massive IT spending on both services and software and hardware. I'll leave the hardware out of this because that's a little bit of a different situation because the price of hardware does keep coming down, but the software prices have gone up a lot. Cost of, I know HCM software has gone up a lot over the years. The cost of security software is massively important and big. The cost of data management software, the cost of front end software tools, the Microsoft stuff is, you know, much more expensive than it used to be. Google, et cetera. Because so much of what we now do at work is automated. We have to buy really good software behind it. So, you know, IT software spending is very healthy. I saw some data this year that said that the IT spend on software and IT and AI related software is up 60 to 70%. So we are in fact as businesses capitulating and saying, yeah, we'll spend more money on tech, that's fine. But what happens is you spend all that money on Salesforce or SAP or Workday or whatever it is and then you look back a few years later and you try to figure out how well it's all working. And sure enough, it's been institutionalized into your company. It's now part of your infrastructure. You can't really get rid of it, but, but you're not sure if you're getting your money's worth out of it anymore because it turned out to be more expensive than you thought. And a bunch of new things are available in the market that the incumbents don't have. So there's a pent up probably never ending discussion about whether we're spending our software money correctly or effectively and what AI is doing.
[00:07:38] It's injecting a sort of big new cost factor into this. Now, I think for most of you, if you're a big company, and you're good at technology. Like, if you're a really savvy technology organization, you have a great CIO and lots of IT people with very successful applications, you don't hesitate to invest in technology. I mean, going back to my early days in my IBM career, the bank of America, which used to be located in San Francisco, and Charles Schwa, which used to be located in San Francisco, were two of our bigger clients. They spent a fortune on technology because they knew that there was no way they were going to grow their financial services businesses without it. And that resulted in, you know, mainframes and fancy software and teller machines and security stuff and, you know, databases, all sorts of things. But a lot of you aren't in that situation. A lot of you have less sophisticated IT teams, you might be smaller, and you don't have the resources to take advantage of all this. So if you do write a big check to Workday or SCP or whoever it may be, you're hoping that it works. And maybe you don't have the technical team or internal operations to make sure that it pays off. And so you do end up, a few years later, questioning whether your spending was adequate. And AI has very much the same dynamics. If you listen to all the stuff that ServiceNow just announced this week. The reason ServiceNow is a good example of this is because they actually formally stated that they plan on doubling their revenues in four and a half years. So they're not being shy at all about the cost of this, and they're selling the compute cost in different tiers. So you don't just buy credits for AI. If you use the AI for more autonomous things, you pay more per token or per credit than you do if you use the AI for simple things. So they're actually giving you premium pricing also, which I think, you know, I actually credit them. I want to thank them for doing that because they're raising the issue. Where everybody wants it to be is it's a financial issue. So we're going to be sitting around, you know, in our HR and IT departments. And by the way, there's no way you're going to do AI and HR without talking to it. This is a combined relationship between the two functions, because all the employee stuff, that's really a big deal. And it's not just hr. It's employee productivity, which we may be working on in HR is interdependent with the technologies and the architectures that IT buys. So these two business functions are getting closer and closer together. And so the IT perspective Which might be, oh yes, we need to spend a whole bunch of money on the new workday tools, the new SAP tools, the new ServiceNow tools, whatever, and we're buy all this stuff for you guys.
[00:10:24] And then they're going to say to you, but by the way, because we spent all that money, you got to cut 30% out of HR, you got to cut 30% of your staff and you're going to say to them, wait a minute, why did you buy all that stuff in advance of us actually knowing how to use it and build a solution we have? So the trillion dollar reengineering happens in fits and starts and it, it happens one solution at a time. And what I've concluded after talking to a lot of companies about this is that the infrastructure probably should not come too early before the solutions. Now, in the old days where I used to work in database and mainframes and stuff, companies bought the technology way in advance of the solutions because there was a very mature market of implementation solutions to deploy. You know, if you bought a mainframe and all the stuff in it, or you bought a bunch of databases or you know, other tools, you were pretty sure you knew what was going to happen because there was a robust experience set of how to build the applications and deploy them. In this case there isn't. So if you go to one of the ServiceNow conferences and you listen to TD bank or CVS Pharmacy talk about this spectacular thing they did, mostly what they're talking about is not a new AI application, it's an IT service delivery automation solution, which is good. I mean, that's all really good. But it isn't the groundbreaking new business application that we sort of hear visions about, because nobody's kind of done that yet. I think the company that has done that well is Amazon. And they did it all themselves. Amazon builds a lot of their own internal systems, so they're way, way down this learning curve. But a lot of, I would say most companies are not very good at that. They don't have the visionary thinking yet. They maybe don't have enough AI expertise. There's a lot of AI fear. So if you do rush out and you spend a bunch of money on these tools, I would be very careful that you have projects that you're pretty sure you're going to do that are going to have ROI lined up and almost ready. Now the alternative is you can wait for all of the AI providers to sell you a solution. And you know, they're working on that in hr. Obviously Workday is working on their agents and SAP is, and ADP is and, and Oracle is and everybody else. And they look great when you look at them. But I think they're very young. I don't think they're poorly designed, but they're young in their maturity. Because over time what a recruiting AI looks like today will be much, much different. And I think recruiting is the best example because it's the most advanced area of hr. If you looked at an AI powered sourcing system or an AI powered interviewing system, it looks great. But that doesn't solve your whole recruiting problem because recruiting is sourcing, interviewing, job architecture, design, it's candidate experience, it's candidate marketing, it's advertising management, it's onboarding, it's all of that combined into feedback. And then there's IO assessment and interview scheduling. I mean, there's just a lot of pieces to it. And the more sophisticated multi functional AI applications in recruiting do all of that. But there's very few of them. Most of them only do piece parts. I just had a long talk actually with a very large parts distributor retail company yesterday about, you know, how immature the market is. So you're going to be involved in not only selecting but building a more integrated solution. So the economics that you're being promised by the vendors is a little bit ahead of the benefits you may see now. You know, like I said, some companies are fine with that. They don't mind getting involved in large technology projects. They welcome it, they enjoy it. This is part of their DNA. I'd say it's a relatively small percentage though, because if you don't have a project staffed or ready or defined, these tools just come along and you come back a couple years later and you say, you know, we bought all this software, we're not using it. And then you blame the vendor, but it's not the vendor's fault, it's our fault. So my recommendation to all of our people that we work with, and mostly hr, but a lot of other people is work on your use cases, work on your business case. Think about where you want to apply AI for the greatest benefit of your company, not just generally, and get a team staffed up on that. And that's by the way, we are very happy to help you with that. That's what our HR 2030 project is all about, is showing you these opportunities and helping you scope these out. And what happens is once you get a project defined, you know, maybe it's global onboarding or maybe it's career development during a merger and you're going to see a lot of this stuff come out in the next two weeks, there's a whole bunch of announcements here. Then you're going to suddenly see there's all sorts of interesting little angles to that project or that solution.
[00:15:31] I talked yesterday, I think, on the podcast about the Microsoft agent for crisis management. There's an announcement that's going to take place on May 20th from a vendor about a big new AI HR system from a company, you know, that's going to be used by Paramount for the merger that they're going through, which is a very complicated process, but AI is going to be a big part of that. So whether you're a parts distributor that's into hiring or you're going through a merger, or you want to completely revitalize your engineering organization, you're trying to do a skills assessment, or you're trying to grow into a new geography and you need to rapidly redeploy people. Those are use cases where you could take your AI knowledge and you could sit down and say, you know, if we could build this, this, this and this, we could do this 10 times faster and 10 times better. The ROI would be $100 million a year. Okay, let's go buy this $20 million thing from ServiceNow and get going. So that's the economics. The other aspect to this that I want to talk about this morning is the people stuff. Now, we're going to do some interesting things at our conference. Irresistible is June 8th through 10th. Uh, it's. It's almost full, but there's some seats left. So please sign up. And we're going to launch a whole bunch of things there. But one of the things I'm really excited about is a new book that I wrote with support from Kathy Enderas that that discusses the future of work and humans in the age of AI. And it isn't a pontification or dreaming book of the future. It's real. It's a bunch of stories, real stories about super workers, super managers, super teams and super organizations. And what you'll hear and read and learn about from. From us, which maybe everybody doesn't quite understand yet, is that the number of people we need in our companies isn't going to shrink down to zero. It's not going to be an 80% reduction. Because what will happen, and I think we're seeing this in the job market right this minute, is in the places where you automate. By the way, health care has a lot of automation, and look at how many new people are being hired in that industry right it hasn't really reduced the number of people needed, it's just. Just improved the scale and scope of the services is that as you automate away the routine work and, you know, things that didn't feel routine before now are routine because the AI can do them, the human value add gets greater. And the example that I think is an easy one to understand here is the old Ritz Carlton story about human judgment. And if you've ever read about the Ritz Carlton, which is now, I think, owned by maybe Marriott, I forget what they used to do. And I think this is just a really good lesson in HR and leadership is they would say to people, we're going to teach you all about your job in the hotels, we're going to teach you about hospitality, we're going to teach you about food, we're going to teach you about drinks and all the things you need to do. But your job is to use your best judgment to do what's right for the guest, because it's the guest experience that's the most important. And there's no way we can design every workflow, quote, unquote, or business process or rule or system to be perfect every time. Because in the hospitality world, everything happens in an unpredictable way. You've been in a hotel, I travel a lot. All sorts of weird things happen. The room's not set up correctly, something broke, the light bulb burned out, the heater's not working correctly, cleaning staff forgot to, you know, put the soap in the dispenser or whatever. And there can be checklists for all this, and of course, and there can be all sorts of rules to try to prevent that from happening. But ultimately we're taking care of people. And I really do believe, you know, in my business experience, that the best companies that I've ever worked with and worked for and been a part of basically are in business to take care of people. That is what we do when we sell people stuff. I know the tech industry is very excited about, you know, pure product for companies that have no services, that are, you know, 99% profitable at a margin level. But that's not really what happens. It may feel that way to, you know, Apple shareholders or Amazon shareholders, but the actual operations of those companies is there's a whole bunch of human stuff going on. I mean, when you go into the Apple Store and talk to the Genius Bar, that's all human stuff. Those are training issues and culture issues and, you know, et cetera. Same thing with Amazon. When they deliver your package, somebody delivered it, somebody Sorted it, somebody picked it off the shelf. And a robot may have done a lot of it, but there were humans behind it. So as we automate more and more things, the human value add becomes a higher percentage of what people do. And that's what's happened in healthcare. You know, the automated X ray imaging machines and scheduling machines. I mean, it's amazing how easy it is to schedule a doctor's appointment, at least in my Stanford thing that I use. But I still have to talk to the doctor and I want the doctor to know who I am and remember what we talked about last time I was there.
[00:20:34] So that human stuff is becoming more important and more valuable and more scarce in every company that automates. So what you're going to hear about in our book and hear about from me and hopefully experience this yourselves, is the number, quote unquote, of people may not go down. It may be fine, it may go up. And that's why the job market isn't collapsing, is all of this. Automation is great, but as you automate away more routine work or software code or whatever, you create opportunities for more value add by humans. And business is, unless you're corrupt, fundamentally, a value creation exercise. If you're a CEO or a CFO or a founder or an entrepreneur, you're trying to figure out how to create something, build something, deliver something that adds value in a way that your customers are willing to pay for it, willing to renew it, willing to promote it, willing to cheerlead for you because they are getting value from you. And if you don't think about your business in the context of value creation and always simplifying and making it better and always making it easier to use and always making it more aligned to the particular customer segments you're going after, then you're not really thinking like a business person. And I think a lot of these tech companies are. Sometimes they really want to just build a company and sell it to make a bunch of money and move on. And that's perfectly fine. That's the way the economy works. But that's not business to me. That's startup Y stuff. That's technology innovation. And it is a piece of the economy, but it's not the long run, sort of enduring part of the economy. So I don't think we're going to run out of people. But the economics of investing ahead of solutions is sitting in front of us right now. I would not be surprised if most of your AI stuff that you buy, whether it be from Microsoft or Anthropic or you know, workday or whoever is going to go up in price fairly soon. And it's not just because of the cost, it's because of the value. It's because of the engineering going in behind it and the uncertainty in their businesses. So we're going to have to be more and more vigilant on locating and finding and aligning and creating these value creation projects so we get a return on investment.
[00:22:55] This is not, to me, going to go the way of the ERP craze for the last couple of decades where we sort of took it on faith that if we bought Salesforce, our Salesforce would operate better and we just bought it and implemented it. This is different because the future use cases are yet to be developed now. And the fun part about this is there's a lot of creativity here. A lot of opportunity for you to think about what really matters to your company and do really cool things. And we have lots and lots of examples of that. And any of you that are doing exciting, cool, innovative things in AI, otherwise in hr, please let us know. We're going to be highlighting and celebrating more than, I think, more than a dozen pacesetters at our conference. Really amazing stories of what companies are doing here. And we all have to sort of look at these things and learn from them because it's all so new. So that's a little bit about what's going on in the economics of this. And as these big companies come public and we see more about what's going on under the COVID it'll be more clear. I hope this little bit of a rambling podcast has been helpful over the weekend for you to think, think it through. I have no question in my mind that this wave is real and we're going to all get value out of it. It's really a question of how you get there from here. Thanks a lot. You guys have a great weekend.
