Workday CEO Carl Eschenbach: Building the System of Record for the AI Era

0:00As soon as you go in and you talk to the enterprise, a CEO, CFO, CIO, whatever it may be, it immediately goes to an ROI conversation. Like, this is going to save me a boatload of money. And in some cases, it is. That being said, if you only focus on the benefit of the ROI, what happens is immediately, if you're an employee, you start to think, man, my job's being replaced. We need to flip that narrative as well. We need to start to think about how do we use this technology to drive ROI, to get dollars to reinvest in the business to drive growth. We need to talk about AI from a growth value proposition, not just an ROI value proposition.

0:47And that's how you get the employees and agents and AI to peacefully co -existing enterprise and really drive growth for companies.

1:11Greetings. Today we're joined by current Workday CEO, former Sequoia partner and perhaps the most beloved executive in technology, Carl Eschenbach. Workday, as you may know, is a system of record for people and the numbers produced by those people. As we move into a world where people are complemented by AI agents, workday is on a path to become the system of record, not just for people, but also AI agents, and of course all the work done and numbers produced by those people and those agents. Carl talks about Workday's three -pronged approach to monetize AI through seat -based pricing, role -based agents, and consumption -based API access.

1:59He also talks about how to implement AI and how to transform a company while preserving your culture and values. We hope you enjoy. Carl, welcome to the show. Thank you for having me. It's great to be here with both of you. I couldn't think of two better podcasters to be with. and Sonya and Pat, come on. We couldn't think of a better venture capitalist who is moonlighting currently as ECO to be doing the show with. Okay, moonlighting, I like that. Actually, I am moonlighting because I have responsibilities here with you and my other job, right? Some days I wonder which I'm doing more. No, I didn't say that.

2:35I'm doing the other one more if I'm honest. But I appreciate still being part of the partnership. It's an incredible opportunity to still engage with what I think is the most iconic venture capital for all times the Koya and just to be around it in some capacity and see people like you and all my partners is truly a blessing for me to still be involved. So thank you. Thank you. We really appreciate that. We're going to start with the question on work day. Okay. So you guys are the ultimate system of record for people. And a world where AI is replacing people.

3:14What happens to We are the system of record for the two most critical assets of any company. They're people and their money. Okay. Because we have the financials platform. Good distinction. Interestingly enough, they're one and the same. And we get asked and I gasped all the time, your system of record in the enterprise, AI is coming. It's a transformative technology. It's going to disrupt everything, especially in the enterprise. companies like you might be in trouble. And I would say with a biased towards being biased, I think we're in a very unique position. And let me explain why. Today, 20 years old as workday, we've been managing and supporting our clients most critical assets, as I said, Pat.

4:05They're people in their money. Yep. And we now have a database, a highly curated database of more than 70 million users that is processing 30 % of the job wrecks in the United States last year. And when you think about AI, at the end of the day, it's all about data. I think you guys would agree with that. Now it's not just about data, it's the context of the data. And when you have that much data and you have the context of it and you're in the actual workflow The business process workflow You can then use agents to drive actual results that people see value in doing in using This is where we start to talk about things like roll -based agents in the impact it can have on the enterprise The other thing I'd say when you have 11 ,000 customers You've been around 20 years You've been that core system of record.

5:07It led to us becoming the system of truth. People trust us. And anytime you go to one of these major tectonic shifts, from one technology to the next generation of technology, I believe, especially in the enterprise, the enterprise customers look to us and say, how are you going to help us with that transition? Just like 20 years ago, you helped us with the transition from on -premises to the cloud for HR and finance. The combination of the data, the context of the data, we're in the workflow, the trust we have, and I think at the end of the day, as we talk about agents, and I'm sure we'll talk about it here, I think domain -specific agents with highly curated data are much more beneficial in the enterprise and just general -purpose LLM models that are out there.

5:59I agree. That's why we think we're uniquely positioned. I get asked all the time, are the incumbents in trouble because of the startups? And quite frankly, it's not an or. It's an end. There's going to be success across the board. You guys have a bunch of great companies that we work with that actually bring more value to the system of record. So it's both. I think there's opportunity for everyone, Pat. And your point on trust is a good one. And that's definitely been a recurring theme amongst the application layer AI companies that seem to be working. Let me ask you a question on product, though.

6:32So let's say that I am an AI developer of some sort or more practically. Sonia is an AI developer of some sort and she shows up at workday because she heard that there's this treasure trove of information that you've amassed over the last 20 years, all of these records and all the context around the records. how much of that is actually available for you to go build cool AI stuff with. And how much of that due to permissions or information architecture or 20 -year -old company stuff, how much of that is not available or said differently? And actually building with AI, all the data that you've amassed is obviously an advantage.

7:14How accessible is it? And what roadblocks are there to making the most of all that data when you try to turn it into product? Yeah, it's a great question. If you look at the amount of companies and technologies that access the Workday System of Record, it's a lot. For example, identity companies. For example, Active Directory. When you onboard a new employee, guess where they all come and do that on top of the Workday platform. Today, they do it through an open set of APIs. General -purpose APIs that we put out there for people to develop on and connect with us. Going forward, as we start to think about this world of AI, in the potential of all of these agents entering the world in the enterprise, someone has to think about how you onboard them in a secure, responsible ethical, compliant way and make sure there are identity and access controls and all the the day they get access to has governance around it.

8:17Today, that's the risk. That is the risk in the enterprise. When we talk to CIOs, the first thing they say to us, we love agents, we love this technology, we love what potentially could happen with agents, but we are concerned we're seeing agents sprawl across the enterprise. My analogy, Pat, is remember when we always used to talk about shadow IT creeping up? And how do you get it back? Because of the security and compliance and control, controls it, you want to incorporate headquarters or a corporate function, that's what we're seeing. Now, what we're seeing is people want to bring agents into the enterprise.

8:52They want access to our records. They want to onboard them just like we would a human worker. They want to do it with a digital worker. And what we're building is something called an AI gateway where you're not going to be able to do that to a generic set of APIs that we have today. You're going to have to come through this agent, if you will, gateway to on board into the enterprise through us. Yeah, same more about this. So people are using workday or will be using workday to on board AI agents analogously to how they might onboard human employees. Bingo. I'm going to hire you as my next sales rep, Pat.

9:28So what we announced, that was a good setup. Needy didn't know it about a month and a half ago. We announced something called an agent system of record. And think of it as the unification of a system of record between human workers and digital workers. Just like human workers get on board in the enterprise today, they come through work day, they get on boarded, they get assigned to whatever organization they're part of, they get their benefits, they get tracked, they get monitored, they get looked at for performance reason. And then you do benchmarking of all your human employees to drive workforce management and capacity planning.

10:03The world of AI comes about, agents come into the enterprise, everyone says they're going to come into the enterprise, okay, how do they get into the enterprise? Who onboards them? What organization are they part of? What access rights do they have? Who controls them? Who measures them? No one today. We think in the future, and as we talk to our customers with this new agent system of record, that unification of human and digital workers on a common platform is to path forward. And then once that's in place, you can start to do workforce management, workforce planning around all of your workforce because it's no longer a workforce of human.

10:41It's a workforce of human and digital workers and that's what we've developed and we're in a process of bringing to market. That's so fascinating. So it's a system of records for your people and your money but your people includes your your human people and now now your agents as well. Exactly. And there's a lot of fear out there. Sonya that these agents as they come into the workforce are going to completely displace all the human workers right and And we're going to, you know, I hear some of my peers go out there and talk about, we're going to see a 20, 30, 40 percent reduction in the human worker right going forward by all of the agents in this whole AI revolution.

11:16I fundamentally don't believe that to be the case. Yeah. It's interesting. I remember the, I remember 15 years ago hearing a Neil articulate the vision for work day. And part of it was the reason people and financials go together and the analogy to the last generation before that, which is the ERP systems of Yesteryear, SAP, and such physical manufacturing was the economy. And so the same system that took stock of all of your machines was the system that was your financial backbone. And with the original workday, we're moving into a knowledge -based economy. And so the same system that takes stock of your people is the system that provides your financial background backbone.

11:58And now as we move into an AI -driven economy, the same system that take stock of your agents is the system that's in your work, financial backbone. And because it's fluid between the people and the agents, it makes sense that we're all living one place. It ought. And if you think about, you know, the workday platform, everyone thinks about us having a HR platform and a finance platform. But if you look at the architecture, they're one and the same. Exactly as you just described. Now layer on top, this new workforce called digital workforce on top of it, that benefit that you've got for the last 20 years continues, but it brings these agents into the enterprise in a much secure way.

12:38Yes. Yes. And then practically speaking, how do you take this to market? How do you roll this out? Do you need some sort of a standard unit of pricing and packaging that an AI agent builder can adhere to? A person is a person. And there are lots of different types of people. but ever's got a date of birth, you know, ever's got a title. What are the common characteristics of AI agents that will allow them to sort of map into this framework that you guys are doing? And how do you price them as well? How do you price them? How do you go to market, right? I remember being in Sequoia. I mean, I back up and that on to your question.

13:17Do you remember when this whole, actually, I was talking to rule off about this last week and everything was going to go PLG? Yeah. Everything's going product -led growth. You don't need sales reps. You don't need, you know, 37 -year sales veterans like myself anymore. You don't need the enterprise. You don't need to top down. I'll tell you, the times are coming back with AI, it's a top -down sale on most cases. It's not product -led growth up to the enterprise. You're selling to the heads of departments or heads of functions or CIO. Yeah. So I think the distribution is coming back to a much more inhuman sales motion.

13:53You know, definitely seeing that. We're absolutely seeing. I think that's a common thread. If you look at the AI, I think that it really works. It's fun because I told rule off when I saw them last week. I'm like, see, I told you, PLG is good, but it's not the panacea that everyone thinks you are still going to need people like me someday. Yes. And they do, right? Yeah. So I just thought I wanted to talk about. As far as pricing goes, there are multiple ways people are pricing AI, and then I'll speak specifically about agents. Let me back up two years ago when co -pilots came to market. A lot of our peers, a lot of my peers and competitors immediately went to market and they basically said to their customers, we now have a co -pilot.

14:34And because we have a co -pilot, we're going to raise our pricing 20%. And we made a decision and I specifically said, I do not want to do that. I don't think it's the right thing. Co -pilots are going to become part of the core foundation of all products. And it should be built in, not bolted on. And it should be the value our customers get and receive by paying us a subscription fee on an annual basis. We have to innovate. Yeah. To be honest, we got beat up quite a bit. Why aren't you doing this? You have code pilots. You should be monetizing. And I said, because it's the right thing to do when we bring true value, right, with our agents in the future.

15:12And we can show you a strong ROI or a value equation, then you'll pay us. Fast forward to today. We price our AI solutions in one of three ways. Number one, we price it based on seats. So we are a seat -based company for the most part. And if we have an agent or an AI technology that provides value to all of the employee population, then we will do an uplift based on that agent or that technology. Uplift per seat. Per seat. The second would be based on an actual agent itself. Like if I can come to you Pat and say we're rolling out a payroll agent and you think about how many payroll administrators you have today but we bring a roll -based agent into the market.

15:58I'm making this up and we're going to say it's $50 ,000 a year for this agent. Today you have a payroll agent or a payroll employee over here looking at everything all you they're doing is administration and that costs $200 ,000 you'll say hmm I'll do this. So you can do it based on the role. Do some of those cannibalize the seat -based model of core workday, meaning if I now have a payroll agent, I need to hire fewer payroll people, and I would have been paying for a seat in workday for each of those payroll people. Does the agent business cannibalize the sort of original seat -based business?

16:32I don't think so. Okay. Because what AI should allow that payroll administrator to do is go off and do more high value tasks and drive more strategic outcomes for the company as opposed to just being replaced. If you think about what we do today, you, me, everybody, we spend the majority of our day working with technology, sitting at our desk, working on a phone, we're working with technology. The power of AI is going to happen when we transform that equation into technology works for us. Yes. And we don't even know what's happening. Right now, we're interfacing with all of these chap bots and all of these different LLMs out there as a consumer.

17:14But in the enterprise, what's going to happen and the value is going to happen when these agents and AI technologies start to work on our behalf? We don't do those tasks and we go off and do other things. So we're going to flip the equation around. So that's where I think the value comes in for existing employees to go get new skills. Just like the AI momentum could be called a revolution. I think we're gonna go through a skills revolution with the existing workforce to train them to go do other things. And if you look at the history of time, people like all of us, all employees, they're very pliable.

17:52They're adaptable. And what do we always figure out to do how to leverage technology. Technology is the single biggest source of productivity gains in the history of time, for sure, right? That leads to GDP growth, at least a company growth. The other thing we need to work on is this narrative that's out there. As soon as you go in and you talk to the enterprise, a CEO, CFO, CIO, whatever it may be, it immediately goes to an ROI conversation. Like, this is going to save me a boatload of money. And in some cases, it is. That being said, if you only focus on the benefit of the ROI, what happens is immediately, if you're an employee, you start to think, man, my job's being replaced.

18:41We need to flip that narrative as well. We need to start to think about how do we use this technology to drive ROI to get dollars to reinvest in the business to drive growth. We need to talk about AI from a growth value proposition, not just an ROI value proposition. And that's how you get the employees and agents and AI to peacefully co -existing enterprise and really drive growth for companies. Sorry, we've sidebar. Let me come back. There's another way to monetize it and how you're thinking about it. Okay, so we've got the first off, we have uplift on all the seats. Second, when we have price per agent.

19:16Yeah, think about that value -based pricing, right? Okay. And then the third is consumption -based. Okay. Meaning how many times is something like I was talking to Brett Taylor as Sierra, right? They do it based on consumption and outcomes, right? And as you know, being on their board, right? We're thinking about the same things. How many times back to the initial question are people hitting us to get access to our data so we can base on consumption? Okay. So, when all of these agents come out in the world, and all of these things, we're going these agents say, wow, I need to get access to the data to be registered or whatever it is.

19:50We got to get access to work day. Okay, great. Please do so. But we're going to charge for that access going forward as opposed to just that open set of public APIs. So in that case, it's a platform business and it's sort of metered almost it is consumption business. And then you can price it. Do you go out? A lot of people today are going out and saying here, by so many, you know, how so much consumption and then as you do it, you burn it down over time, right? That's a big model it's out there. And we're going to do all the above. You mentioned AI for productivity. Okay. You are easily top three most productive human beings I've ever met, possibly top one, probably top one, honestly.

20:28We were certainly top three. As an insanely productive human being, how are you using AI to make yourself more productive? Have you, do you have any favorite products? Do you have any favorite workflows? Like what? Stereo -gibli pictures. Yeah, any great studio, weekly pictures. Yeah, what are you doing with AI to make yourself more productive? Yeah, so inside of work day, and then I'll talk about how I use it to. Inside of work day, we've rolled out a whole bunch of productivity tools around AI. We have Slack AI, we have Zoom AI, we use Gemini all the time. Gemini is something I use all the time.

20:58So, an example I'm preparing for earnings every 90 days, which is so much fun. Please join me if you guys want. All the analyst reports come out, right? And you know, how do you summarize them? How do you, it's amazing. You throw it in Gemini and boom, and they give you, you save, you know, hours of reading. That's one example where you're on a Zoom call that you missed. Like how does that get summarized back to you? I use those type of things all of the time. Yeah. And it just makes me that much more productive, clearly. One of the things we're doing inside of Workday, and I think it's first of its kind at scale.

21:35We're launching this month something called everyday AI. We're all 20 ,000 employees have to go through training on AI and how to learn how to use all of these tools to drive personal productivity and do that personal productivity gain work they get gained. Yeah, yeah. Right. So we and they have to test out. There's an entire curriculum. It's a really interesting way to get people excited about AI. I had to engage with it and leverage the technology as opposed to be afraid of it. And we think it's one way to drive it, if you will, that peaceful coexistence between the technology and the humans to drive overall gains for the company.

22:16And that's another example of how the whole company has to do it. And I'm launching it in two weeks. Very cool. Can you tell us more about the agents that you're bringing to market? So you mentioned the payroll agent. What other agents should folks expect from you? We bought a company at this time last year about a year ago called HiredScore. And they have what I would describe as a recruiter agent. So if you think about recruiting Sonya in the HR function, it is town acquisition or recruiting is probably one of the most expensive functions you have in HR. This recruiter agent we brought to market, we can give it to a recruiting organization or town acquisition organization, and they, in a very short matter of time, can see a 50 % increase in recruiter productivity.

23:04Just think you get hundreds of resumes for a job. Today, someone's searching, now it basically can do that in seconds and say, here's the top two or three you should go after. But it's more than just recruiter productivity. It also has proved to accelerate time to higher by 20 -30 up to 40%. So your recruiters are more productive and you can accelerate your time to higher. Then once you get them inside the building, we have two other AI products. One is called talent optimization and the other is a talent mobility agent. And when you implement them, we have seen upwards of 40 % reduction in attrition.

23:48And the reason for that is for the employee, it takes a look at all your skills, your background and your experience. And for the manager, you say, I have this project or outcome I'm trying to achieve. And it starts to do the matching between the two. So what it drives is an increased level of internal mobility with your workforce, which drives down a trition because people feel like they're getting new opportunities. So that's like a full life cycle of agents from recruiting to onboarding, to town optimization, to talent mobility. And all that we can monetize because people can actually see, you see a reduction of 50 % of, you know, you know, need for additional recruiters.

24:30That's pretty powerful. Yes. Can I ask you a question on, you mentioned this idea of matching talent to other opportunities inside of their company. Yeah. One of the ideas that we've heard from a few different companies at this point is this idea that thanks to AI, the matching problem that is, here is a job to be done, here is the right person to do the job. Thanks to AI, that matching problem is now possible. Historically, it wasn't because you'd have to collapse the dimensionality of that problem to things like what is the title, what is the geo, what is the pay scale, which really doesn't tell you whether or not this person is going to do a good job.

25:05And so I'm curious, are you a believer in the ability of AI to actually match the right talent to the right opportunity? And if so, is that a big theme for Workday going forward? We heard a glimpse of it with the internal product. Is there more of that to come? Absolutely. And it all becomes completely or highly dependent on skills. It's all about skills. It's not about pedigree or college or anything else. like, you know, they're, let me give you a one statistic. Um, and then I'll come back to say how we do it. Accenture, one of our biggest partners out there, who's a very reputable as you know, system integrator, global system integrator.

25:47Today, their new hires 30 to 35 % of their new hires do not have a college degree. Hmm. It's pretty amazing because what they've done, they're probably one of the most advanced companies I've worked with that looks at skills and job outcomes or job wrecks are trying to fill and then they use AI to match the till. Right? And that's what our product does internally. That's so you have 10 ,000 employees because we are the system of record. We know all the skills you have and you have this if you will database of all the skills of your employees. So now when you want to try to hire someone, you first look inside and say, this is what I'm trying to fill this job.

26:28These are the projects this person's going to work on. And these are the type of skills that are required. Then you look at your own 10 ,000 employee population and say, wow, here's all the skills Billy or Susie have and you do that internal matching. And so as we move to a skills -based outcome world, right, this becomes more and more valuable inside of companies for people not to just hire based on pedigree is how we like to think about it. Carl, I remember when I interviewed Pricycoia, you were very memorable. I mean, so much of so much of that interaction though, it was just the human. Can I see myself working with this person being inspired by being around them every day?