Episode Transcript
[00:00:07] Speaker A: Good morning, everyone. Today is the weekend before workday rising. And so I would like to talk about the workday company a little bit in the context of a really controversial blog post and announcement by Klarna that they were going to replace workday with their AI tools. Klarna is a buy now, pay later company located in Sweden, I believe about 5000 employees in that range. So they're a mid sized but not small company filled with software engineers. And the announcement they made, the concept that they're endorsing is to replace Salesforce and workday with their own AI systems. And the assumption is that they can build the equivalent of workday and Salesforce through generative AI because generative AI could essentially read the documentation of these systems or the APIs and create the workflow and business logic and database structures needed to operate these two systems. Now I'm not going to talk about Salesforce, but I know a lot about workday.
This is a, this is not the first time this has happened. Google built a lot of their own HR software years ago. GE did. I mean, I've talked to many, many big companies with large it shops, Walmart, et cetera, who have built a lot of their own HR software later to conclude that it was a bad idea because it was very expensive to build, very expensive to maintain, and almost impossible to keep up with the trends and directions in human resources. And to say nothing of the fact that the engineers who work on the core HR system are not working on the product that Klarna sells to customers. So you're taking a cost structure of maybe 200, 300, 400, $500 a year per user that you're spending on workday and you're taking that money and you're using it to fund and internal engineers or maybe new engineers that could have worked on the Klarna product to replicate something that's essentially almost a commodity in the market. But they kind of missed the fact that workday has hundreds and hundreds of years of experience in intellectual property and knowledge and product management and understanding of recruiting practices and payroll practices and global employment practices and diversity practices and training practices, et cetera, et cetera, et cetera that Klarna knows nothing about. Now two examples to highlight the danger of this. The first is Google. Back when Laszlo Bach was the chro of Google and Google was really on a tear, they decided to build their own HR software. And so they built a whole team. It's quite large. I don't know the number of engineers, but it's not an insignificant number of people that build HR software for Google employees because they don't want people to log into workday. They want them to log into the Google systems. This includes training, learning, career performance management, compensation, succession management, all of that stuff. What they ended up doing was after a lot of work, they did buy workday, by the way, because the core functionality of workday was needed. And they built this large team of engineers to build talent and user interface employee facing tools on top. Now, I've talked to this team at least six times and they have really, really struggled, not because they aren't good engineers, but first of all, they're not the core engineers in Google. So they don't get the resources that the search engine and AI people get. Second, they can't keep up with the innovations taking place in the external market because they're not in the external market. They're kind of doing product management on their own. So every time I talk to them and they show me what they're doing, I scratch my head and think, gee, I don't think you guys are paying attention to what's going on in the space. And they've eventually, just recently decided they're going to go with third party products. Now this has taken a long time for Google to come to this conclusion, but it's a good business decision. The HR software industry is a very sophisticated industry of very savvy smarteen product managers and software engineers that understand these complex data, AI recruiting, learning, workflow, global employment issues that HR has to deal with. And it's not just HR. These are business issues of how do you do business in different countries and how do you pay people and what should your benefits be and should you have flexible work all, every time, every year. By the way, there's some new work trend that comes out, whether it be hybrid work or four day workweek or gender equality or whatever it is, you got to fix your HR software all the time. In fact, I believe there are something like 15 or 20 genders now in workday. And I don't know if the average person at Klarna would have thought of that when they were building their own thing. So I really think this is a, maybe it's just a pr stunt, but it does beg the question of a bigger issue, which is where will AI fit in the disruption of HR software? And for those of you that read the article, why I hate workday, by the way, I'm not paying. I'm not trying to focus on workday. Oracle successFactors ceridian, which is now called day force ADP everybody has these challenges. These are very, very complex systems and they're often hard to use just because they have so much functionality for so many different types of users. That's why they are generally very large teams working on them. So where does AI fit? Well, there's going to be a lot of announcements in the next 30 to 45 days. I don't think I've seen all of them, but I've probably seen a lot of them. And generally speaking, I would say there's one massive change in addition to talent intelligence, which changes the operating dynamics of these systems and makes them more personable, personalizable, easy to use in all these different talent practices, and then these buttons that can generate things like job descriptions and other things like that. There's a bigger, to me, issue, and that is what is called agentic AI, which I wrote about last week. Take a tool like Galileo or chat GPT, or now assist from Servicenow or Microsoft Copilot, by the way, each of which are different in their architectural and data foundations. But, and you stick it on top of this complicated, highly functional HR system, and you can create a system where an individual employee, or a manager, or an HR professional, or a comp specialist, or an HR business partner or a call center agent can ask the system a question, get an answer, and tell the system to do something that is 100 times easier than clicking around through menus and tabs trying to find where to do this and how to do it. Most of these hrtainous transactions are complex. Under the covers, there's security things that have to be checked. Sometimes there's multiple steps. You can't open a job rack until you create a job, and you can't create a job until you create a job level and a job description and a location and all sorts of other things that you may not think about as a user. And the AI can prompt you through all of those questions or decisions and get you to the point where you do something without having to read the book or ask somebody from it to help you. This is a massive, massive change. SAP is pioneering this with Joule, which is designed to do this, and it does it today. It works. It's amazingly functional, and they're rocketing ahead with that work. The workday assistant is moving in this direction, the ADP assistant is very far along in this direction, and ADP is going to make some big announcements in a few weeks that you'll be hearing about. And Oracle, I'm sure, has a similar assistant. In fact, Oracle announced 50 AI agents at their user conference this week, about a third of which are in the HR HCM suite. Lattice is working on this, as are many, many others. So the big sort of change in HR technology that's going to happen over the next couple of years is this idea that the user interface is sitting on top of the system in a conversational platform that looks more like a chatbot and less like a bunch of panels and menus. Now that's not to say that this conversational agentic AI thing is a core application. It's not. So this concept at Klarna, that they're going to use these gen tools to build the whole application is just not reasonable to me.
And by the way, building agentic AI in HR is not going to be a quick process. The best example I can think of is paradox. Paradox is an eight or 9789 year old company, very, very smart, very sophisticated founders and engineers, that tried to do this simply for candidates, job candidates. And they found that when job candidates ask questions about jobs, they ask things that require access to the back end systems. What are the hours? What jobs are available in my city? What's the pay range? What's the culture like? Do you have family benefits? Do you have benefits for pets? Blah, blah, blah. That is not going to be written down and hard copy in the chatbot. That has to go into some system to find it. So what paradox has learned over the last six, seven, eight years is how to build these conversational interfaces into backend systems. It's not something they did in a year. It took a while. Now, at this point, Paradox is a total recruiting platform that includes candidate experience, interview scheduling, assessment, applicant tracking, and many, many other things. And now it's incredibly powerful. It is an application. And actually, to some degree, Paradox did build an ATS, and so now they are in fact, to some degree competing with these larger backend systems. That is not to say that everybody's going to do this. Servicenow is not going to build an ERp. They're going to build all sorts of other tools, but they're going to assume you have an ERP. ServiceNow, by the way, has built a lot of functionality in employee growth and development and training and career, which could compete with some of the platforms you have. So there will be a slow process of the new AI platforms eating into the backend systems. But don't be fooled. The backend system vendors know this very, very well, so they're going to build their own too, or partner with vendors that can help them. And this idea that the front end is going to replace the backend is just nuts. I mean I worked, I spent a lot of my career at the mainframe industry back at IBM, and those mainframes are still there. They are behind the covers running your bank. Most banks have a significant amount of their transaction processing on mainframes. That's 50 years later. They're still there because the business rules, the transactions, the scale, the security that's embedded in all of the years of experience in those systems cannot easily be replaced. So this is a, you know, architectural transition we're going through. And the agentic AI platforms, by the way, haven't been totally figured out. You know, we went through the workflow options for the copilot several times with Microsoft and they admitted to us that the copilot studio is a work in process. They haven't decided all of the APIs and how they're going to work relative to back end systems. Sana, who we work with, is doing the same thing. So we can figure out how to do our integrations with back end systems, as will, you know, the other vendors servicenow as well. So, so we're in the sort of a transition phase there will be, by the way, in my opinion, and I'm going to talk about this at HR tech and at unleash and to some degree I'll talk about it at workday rising. There will be a new generation of HCM. It's going to come slowly. It's very slow and very hard to replace large enterprise systems. I mean a lot of companies still have Peoplesoft modules kicking around, but you know, some of the really old ones are gone and they eventually do get replaced. My experience is that about every decade, every seven to ten years, a large company will go through a sole searching process of deciding if their core systems are doing what they want or if they want to upgrade them or replace them. And usually they have to issue an RFP and that's where new platforms get in. And by the way, the other interesting thing about the core HR market is a lot of the next gen vendors, ADP, Lattice, Highbob, Dayforce, have aspirations to be the next workday. That's a wonderful aspiration. They will probably get there in some respects in the mid market. But my experience, again, going back to the mainframe analogy, is that these large enterprise systems tend to stick. And you know, if you think about a company like Microsoft, which is a trillion dollar plus market cap company, there is no financial value for them to switch HR platforms. Even if they're frustrated at some feature and success factors, they are partnering in. Actually, they do the opposite. They are partnering with SAP to help SAP build a better interface into the new Microsoft front end tools. That's the way you need to think about it. If you're a large company and you decide to do business with a vendor like workday, you're a partner with them and you're paying them a relatively small fee for a massive amount of R and D they're doing on your behalf. If you reach a point, of course, where it just doesn't do what you need, then it's time to really either make a stink or look around. But, and of course, vendors have to accommodate hundreds and hundreds of customers, so they can't do everything that everybody wants. But, you know, this idea that you're going to replace these things with AI, it's just not going to happen. One more example or analogy of this. The iPhone 16 launch this week from Apple has been very unimpressive to a lot of people, including me. We believed that Apple was going to give us a Siri that would replace kayak, opentable, all these consumer apps that would be able to book a restaurant, order things delivered, go into Amazon or wherever and get things, schedule a flight, go into our outlook calendars, et cetera. They haven't announced any of that, will they? Maybe. I mean, hopefully they do. That's really the perfect agentic system is our phones. But, you know, it's hard, and I think one of the largest, most successful companies in the world, Apple and a very, very successful software company is struggling. So, you know, they're not going to say that, but that's just my opinion. So just to sort of make the point that this is not a trivial effort to replicate and replace these back end systems that we've lived with for a decade or longer. Anyway, two more things. First of all, I'm going to be keynoting and doing a whole bunch of stuff at HR Tech in two weeks in Vegas. I will be at workday rising. Next week. We're going to be launching our partnership with Workday there with Galileo. We have some Galileo demos for those of you that are going to workday several weeks later, I'll be at unleash and then I will be at workday rising in Europe. So I have a lot of travel going on. But there's really two things I want to highlight. Number one, AI is here and it is really happening. And you're going to see some announcements from us and some research from us to show you great examples of where AI is being used. The second is you need to learn about this stuff yourself. I can't tell you everything that you need to know. One of the things we'll be introducing is a new version of Galileo that will make it even easier for all of you to get access to it. Something we call the Galileo success center and more education and training on AI. Because all of us are going to be using this stuff. We can't be intimidated by it. We did a webinar, Bill and I did a webinar yesterday, and something like 70% to 80% of the people on the webinar said that the obstacle they have to these AI platforms is a lack of understanding and lack of clarity of how they're going to use them. I think that's going to go away very, very quickly. And for those of you in the HR tech market, I'm sure you're excited by all of the innovations you've come up with. There are a lot of innovative things I've been seeing to get ready for these conferences, and you're going to hear about them from us, and you can always call us to talk about them. So I really want to tell you how excited we are about the direction we're going in HR tech. So that's a little bit of perspective on where we are with the agentic AI and AI chat and interfaces in the human capital market. And we look for. I look forward to seeing a lot of you at these conferences. Just come up and talk to us. We'll be all over the place. We have a lot of people coming to them and lots more to come. Thanks a lot. Have a great weekend.
[00:17:40] Speaker B: Sadeena. Sadeena.
