Episode Transcript
[00:00:00] Okay, guys, today I want to talk about Workday's earnings and the SaaS apocalypse narrative that seems to be changing.
[00:00:10] So the story for the last three years that most of us have heard from the financial analysts and the rest of the world is old SaaS software is out, new AI software is in.
[00:00:22] AI is going to replace SaaS. The AI companies are growing at 100 times the speed of the traditional software companies.
[00:00:31] So Salesforce, HubSpot, Workday, Oracle, SAP, Adobe, all of them are yesterday's news.
[00:00:40] Let's move on and do something more exciting and build everything from scratch in AI.
[00:00:47] Well, that's not exactly the way the world works.
[00:00:52] Two things that are in the way. Number one, the AI software companies, whether you call them the Frontier Labs or the others, really don't have that much real end user revenue yet from corporations. The corporate IT software industry is about $1.4 trillion, but most of the revenue coming to Anthropic and OpenAI is coming from other software companies using their tools to build other software.
[00:01:18] So the actual AI spend, which is going up a lot, is not $1.4 trillion on AI, it's probably 1.1, 1.2 on SaaS and maybe a billion or two at most on AI.
[00:01:37] So these big legacy systems like Salesforce, Oracle, SAP, Workday are still the software that runs our company.
[00:01:47] These are the systems that pay people, that hire people, that keep track of our record keeping, that train people that understands your job role, your skills, your pay, your rewards, your benefits, your vacation time, the hours, the scheduling of your hours.
[00:02:04] All of that stuff is being done by these SaaS systems.
[00:02:10] And this is mission critical because if any of these things break, your company doesn't work right.
[00:02:16] And my experience doing this for close to three decades is that even if your vendor, your HCM vendor is really a mess, it's really hard to change vendors. And so companies don't replace their HCM very often, if ever, unless the vendor you're working with goes out of business or is acquired and just doesn't develop their product anymore. And that does happen every now and then, but it hasn't happened for a while at least.
[00:02:45] I mean, I suppose there's people on PeopleSoft that, you know, a few left that are trying to figure out what to do, but most of them have been migrating to have already migrated to Oracle or Workday.
[00:02:57] So anyway, so we got this big, big market of people running their companies on these, you might call them legacy systems. We used to call legacy systems systems that work once they worked. Our legacy and then this new stuff comes along and everybody goes, oh, it's so great. Let's build a whole new HCM platform out of agents.
[00:03:15] Well, that is not what's happening. And the Workday actually announcement this quarter, their financials actually points this out because the capabilities and features that are introduced in Agentic HR, which we call HR 2030, do not replace everything you're doing in your current HCM system. They are additive and complementary and they improve and speed up and accelerate your ability to build new applications and to share data and to analyze data and to use data, but they don't replace everything you have.
[00:03:52] And so even though it took a while for the HCM vendors to figure this out, and of course, you know, in the early days of AI, no one really even knew what it was.
[00:04:01] So there were lots of false starts by all the incumbent vendors on what they were going to do.
[00:04:06] We now know that you could take a platform like Workday, which is very unique because it's a single line line of code for everybody who runs Workday. Everybody who runs Workday uses the same release of Workday, which is not true in Oracle or SAP Land at all. And Workday has an object oriented database and a graph and all sorts of, you know, interesting security layers built in that actually Workday is a really good platform for AI because if you build the right stuff on top of it and you have the right context, all of the data in Workday can be leveraged to either extend or protect the AI agents and applications that you build.
[00:04:48] And that's exactly what Workday is doing now. They acquired Paradox, they acquired Hired Score, they've acquired a bunch of other companies. They acquired Sana, of course, and so they had to buy some of this technology to get it off the ground. And of course they've been building a lot of their own AI, going back to machine learning days for a long time.
[00:05:08] But now we're seeing in the earnings announcement that came out last week, if you listen to it, a very confident, clearly articulated, high value, successful strategy for AI. By the way, I think the company that didn't do this very well yet is Salesforce. Salesforce has been rather schizophrenic on their messaging and I don't know that all of the promises they've made have come true, but they're in the same boat, as is HubSpot, as is Adobe, as is everybody else. We also, by the way, we use HubSpot in our company and I've seen the AI stuff that HubSpot's doing. And it's pretty, pretty amazing. It's pretty good. So getting back to Workday.
[00:05:51] So when Aneel introduced the quarterly call yesterday or the last week Q3, 2027 as our number, he made the point that Workday is not an ERP or just a financial and HCM applications company. It is an enterprise context platform.
[00:06:11] In other words, AI agents need to be grounded in real information, real data, relationships, policies, permissions, transactions structures, security rules. That's already in Workday.
[00:06:26] So in order for AI to Workday, we need to connect it to Workday anyway. If you think you're going to throw Workday out, where's all that information going to go? You need to put it somewhere.
[00:06:37] I suppose you could put it in databricks. That's what databricks would like you to do. But that isn't happening anytime soon.
[00:06:43] So in some ways it's Workday's single data model, security model, business process framework, cloud architecture is perfectly suited for AI.
[00:06:54] And they're now articulating this very clearly. They call it lawful agents versus lawless agents. I'm not particularly crazy of that way of describing it, but you get the message. And so what they've now done is they've proven this with the implementation of sana. Now, SANA was acquired last year. We've been working with SANA for quite for at least three years, maybe it's more like four years.
[00:07:19] And it is a very powerful AI AI company, not a huge company, but they're very smart people and they've got a lot of application experience in the learning space in particular.
[00:07:30] And what they've now done in the SANA experience is virtually replacing the entire Workday user interface. I think that's going to be formally announced in October at Workday Rising.
[00:07:42] That's what they said. And that means all this frustration and complaining about workdays too hard to use and I can't find anything and all that gone over, no more problems. SANA is very easy to use, it's very powerful, and you can build agents yourself. Workday announced that they deployed what is called SANA Enterprise, which is the enterprise version of SANA, within Workday. They now have, I don't know, 15,000 employees using it inside of Workday, and they built 22,000 custom agents in three weeks.
[00:08:15] An agent is essentially what an individual would do, like with a spreadsheet, building your own tool on top of the system Workday. And the nice thing about this strategy is if you're using SANA in SANA Enterprise to build an agent that, say, you know, checks your emails or manages your pay of your team or looks at sales quotas or whatever it does anywhere that it touches data that's related to the workday system, financial or HR data. Workday protects you from getting access to data you're not supposed to get, or updating data or changing data that you're not supposed to change.
[00:08:50] So it's a naturally extensible strategy. Now, I'm not saying you can't do all these wonderful things with the Microsoft Copilot or Cloud or OpenAI. You can, and we use the COPILOT internally. And I think the Copilot is going to be one of the most popular ways to do this. But you got to make sure that the copilot is connected to Workday or whatever HCM system you have and that it's inheriting the privileges and rules that you already have in your company. With sana, you get that out of the box.
[00:09:18] So I don't think Workday is trying to position SANA as a general purpose agent for everything you do in your life, but it's going to radically change the opportunity to add value from the Workday platform. And I think most of you guys that use it are going to run all sorts of interesting scenarios and build new things that you could never build before, because now you don't have to be a software engineer or a very, very technical consultant to build something.
[00:09:44] They've also announced that the SANA Learn and the Workday Learning have been integrated, which is great. That's a huge market. Workday has, I think, a thousand or more customers running Workday Learning. I think it's many thousands now. And so I think virtually all of them will implement sana. I don't know why they wouldn't. That's a big business. They announced the talent acquisition technologies from Paradox and Hired Score are now operating with 30 million job candidates and did 8 million interviews last quarter. We know how well Paradox works, and it's getting more and more integrated into Workday every day. And then they announced a new version of the adaptive planning system called Adaptive Decision Intelligence that I have never seen it, but it sounds pretty cool. They have 170 customers who already bought it and it does scenario modeling and sensitivity analysis and regression and planning, and that will be integrated into the HCM over time.
[00:10:41] So the HR 2030 architecture, which we've talked to a lot of you about, will be very, very much powered and fueled by this adaptive decision intelligence tool, which can relate human capital data to financial data, which by the way, nobody really has done very well, even though a lot of people have talked about it. And then in terms of revenue, the numbers are pretty good. I mean, this is not a company that's doubling every quarter, but they basically explained that the AI new ARR in Q2, which was the last quarter, was $600 million.
[00:11:16] It's $100 million of new ACV or recurring revenue assigned in the quarter and it's growing at 200% quarter over quarter. 25% of their new recurring revenue came from AI. So I think this company is in really good shape. And I think it's clear to me at least that Workday is not going away. Workday is not going to be hurt by AI. Workday is going to do a great job of leveraging the opportunities for AI and the financial and business benefits are going to be great for workday customers and workday investors. Now, I don't want to spend a lot of time on SAP and Oracle right now. I think their story is a little more muddled in some respects. But what this also is demonstrating to me, and hopefully to you, is that we're not going to be in a massive replacement market for new AI solutions that replace all the things we already have. And by the way, this is probably going to be true in CRM as well. These applications like Workday and Salesforce are very important, mission critical systems in our companies. And the vendors know very, very well how difficult they are to use and how complex they are to add value and, and they've been suffering through this themselves and they see this as an opportunity to radically improve the user experience, the ability for new code to be developed and for you as a customer to interact and use and leverage the system that you're paying for. If you could build an agent in Workday that did onboarding for your team in 15 minutes using all the data you have in your team meetings and team documentation and SharePoint, why wouldn't you do that with Workday as opposed to buying a whole new system to do that? That's the kind of stuff you can do with Workday and sauna yourself. You don't need to hire a software engineer or a designer to do it. And you can get it out in minutes or an hour as opposed to weeks or months. So the number of applications that you can build in an AI enabled platform like Sana, Workday, or whatever one you decide to buy is really spectacular. And we're going to continue to work on this and share more of this with you guys over time. Anyway, you can draw your own conclusions as you shop around, but I think workday's in pretty good shape.
