Overcoming Data Security and Privacy Challenges with Dave Sohigian, CTO at Large at Workday

[00:00:08] [SPEAKER_00]: Hey everybody, welcome back to the Beat Podcast here at Vive in Los Angeles. Today I am so excited to be kicking off the podcast series with Dave Sohigian. He is the CTO at Large at Workday and tasked with helping customers to get the most out of Workday. He's part of the global office of the CTO responsible for communicating Workday's strategy to customers and prospects and informing the development organization about market needs.

[00:00:38] [SPEAKER_00]: To influence future innovation, Dave joined Workday 14 years ago and most recently served as a global CTO. Dave, thanks so much for joining us today.

[00:00:47] [SPEAKER_00]: Thanks for having me, Saul.

[00:00:48] [SPEAKER_00]: Super, super happy that you're here. And one of the things that really kind of want to just kick off the series with is how is the AI hype cycle similar to previous hype cycles and how is it different?

[00:01:01] [SPEAKER_01]: Well, I've been in technology for long enough to have a fair bit of gray hair. So I've seen more than one hype cycle over that time. I think it's similar to some other hype cycles. I would say mobile technology, for example, or the internet, and maybe a little less similar to some others like blockchain, for example.

[00:01:20] [SPEAKER_01]: And I think part of that is I'm an optimist about this one. I think it's going to be successful as a hype cycle, meaning like this is technology that will positively and definitely change, have a big impact on our lives.

[00:01:32] [SPEAKER_01]: But it's also similar to all hype cycles, even ones that end up in failures, meaning things that aren't useful in the long term, in that people's expectations are definitely overblown in some areas.

[00:01:45] [SPEAKER_01]: Our head of products, Cheyenne Chakraborty, likes saying that AI and ML are overhyped in the short term and underhyped in the long term. And I think it's a good way to look at it. And it's sort of parsing out what are the pieces that are really going to have that long term impact.

[00:02:01] [SPEAKER_01]: And what are you most excited about in the short term?

[00:02:04] [SPEAKER_01]: I'll start with Workday's excitement about it. Workday's been doing machine learning and AI for over a decade now.

[00:02:10] [SPEAKER_01]: One of the challenges we've faced is because we provide solutions to big companies. We call it enterprise. That's the corporate bull speak term for large companies.

[00:02:20] [SPEAKER_01]: Because of that, they're very cautious about their data, especially the type of data we hold, their people and their money data.

[00:02:26] [SPEAKER_01]: So when we propose these solutions over the last decade, often the response was, that sounds great for some other company, right?

[00:02:36] [SPEAKER_01]: Like sharing your data.

[00:02:38] [SPEAKER_01]: Yeah.

[00:02:38] [SPEAKER_01]: We've built up a tremendous amount of trust over the last decade. And I can quote some numbers in terms of how that trust appears for us.

[00:02:46] [SPEAKER_01]: But particularly over the last 18 months, because companies were saying, hey, where's your genitive AI chatbot?

[00:02:54] [SPEAKER_01]: And our first response was, you have to be willing to share your data with Workday, at least in some form, for us to be able to make any sense of it, for us to be able to do anything with it.

[00:03:05] [SPEAKER_01]: So to date, 92% of our customers have signed agreements to be able to share their data. It doesn't mean all of them are sharing it.

[00:03:12] [SPEAKER_01]: Sure.

[00:03:12] [SPEAKER_01]: But those customers have put their trust in Workday, that they're confident, in a legal sense at least, that they're willing to share that data. And of course, I'm sure we'll talk about data as everything when it comes to AI and ML. And it really does determine the winners and the losers.

[00:03:26] [SPEAKER_00]: Thank you so much for that. And so, Dave, it is exciting. And so, Workday, how are you guys thinking about applying AI in its applications?

[00:03:34] [SPEAKER_00]: I mean, exciting to hear such a high percentage of customers have said, yes, we want you to have our data. How are you going to use it in your applications?

[00:03:42] [SPEAKER_01]: Well, this also goes back to our motivations and the incentives for Workday. I always talk about that. Even though I'm a technologist, I understand that in some cases it's follow the money.

[00:03:51] [SPEAKER_01]: Yeah.

[00:03:52] [SPEAKER_01]: And Workday gets paid for providing solutions, software services that provide benefit to our customers in very specific areas, in human resources, in financials, general ledger, in payroll. So that's where we make our money.

[00:04:09] [SPEAKER_01]: We don't sell technology directly. We sell solutions that are based on a technology. So we are a technology company.

[00:04:17] [SPEAKER_01]: So you can't buy an AI thing from Workday. You can buy a solution for understanding your talent, for example, that uses AI, but you're not buying the AI directly. And so that creates an incentive on our part to look really carefully at what tools are we going to use to help solve problems? What are the most effective?

[00:04:38] [SPEAKER_01]: I think going back to blockchain, I don't mean to beat up on it too much, but one of the challenges that blockchain faced outside of Bitcoin was during the hype cycle for blockchain, technologists would say, oh, you could use blockchain to solve that problem.

[00:04:51] [SPEAKER_01]: And the response from older technologists was, yeah, we could also just use a database.

[00:04:57] [SPEAKER_01]: Yeah.

[00:04:58] [SPEAKER_01]: Right? So like, and it was really getting the tool ahead of the problem. And because we don't sell the tools, we're not really tied into, is this something that you solve with analytics or a particular tool or AI or ML, like which one makes the most sense?

[00:05:13] [SPEAKER_01]: And of course, it gets deeper than that when you look at maturing technologies like AI and ML.

[00:05:19] [SPEAKER_00]: That's great, Dave. Yeah. And it's just a reminder that it's possible to overdo it. And I love that you guys are hyper-focused on solutions. And there's a ton of tools that you could use to get there, but it doesn't necessarily mean you have to use one or the other.

[00:05:33] [SPEAKER_01]: Yeah. And I think experimentation, that's fantastic. But when you're talking about, again, big companies trying to make big decisions that impact their future, you've got to also give them really practical solutions to their problems.

[00:05:46] [SPEAKER_00]: Amen to that. And by the way, you know, the best technologists I've seen are also some of the greatest business partners. So love that you went there with that.

[00:05:54] [SPEAKER_00]: How do you see digital transformation reshaping healthcare delivery in 2024? And what are the main challenges and opportunities this presents?

[00:06:02] [SPEAKER_01]: Yeah. Digital transformation has been a huge topic for some time. I was the global CTO before that. I was the CTO for EMEA, and this was back 2017 to 2019. And digital transformation was something they talked about all the time in Europe. And I don't mean to disparage anybody over there, but it was something more talked about.

[00:06:20] [SPEAKER_01]: Sure.

[00:06:20] [SPEAKER_01]: There was chief digital officers there, and I always ask, what's their job? Well, they described something that was like pretty much what a CIO should be doing in the US.

[00:06:29] [SPEAKER_01]: Yeah, yeah.

[00:06:29] [SPEAKER_01]: And that's changed over there. So I think the digital transformation to me is what's behind it, what's driving that. And certainly sometimes it's often fear, fear of competition. You know, if you think of in the consumer space, taxis versus Uber, that's driving the digital transformation of taxis.

[00:06:48] [SPEAKER_01]: And same for newspapers versus Craigslist and other like that. If that's the driver, I'm not saying that's a bad motivation, but you have to step back and look at what's driving it.

[00:06:57] [SPEAKER_01]: And I think the best decisions are generally not made long-term out of fear. It has to be out of what's the goal you're trying to achieve. And if you're looking at in the payer space, if you're trying to benefit patients and consumers, and that's your long-term goal, and you say, we need to be more digital in how we do things to be successful at that, long-term, you're going to have a much better outcome than, oh no, this other provider or this other payer is getting the edge on us.

[00:07:26] [SPEAKER_01]: We need to adopt this new cool technology that we found out about and move as fast as we can. I mean, moving to the cloud, which is a basic premise of Workday, when we first started 19 years ago, we built it in the cloud.

[00:07:38] [SPEAKER_01]: There was no other option. There still isn't. There's no hybrid option. There's no on-premise option. There's no VPN tunnel option. We built it that way right from the start.

[00:07:47] [SPEAKER_01]: But that alone, that's just a technology choice. It's, well, what are you trying to accomplish? Oh, we want to move to the cloud. Well, why?

[00:07:55] [SPEAKER_01]: What are you trying to do? What's the benefit you're going to see from that? And we certainly can point out some for Workday that our customers have had, but that's not necessarily a fit for everybody.

[00:08:04] [SPEAKER_01]: Like that may not, you need to outline those goals and processes you want to change first.

[00:08:08] [SPEAKER_00]: I love that. And at the end of the day, it's what is your strategy? And what are the levers that you want to move? And what are the KPIs and metrics that you're looking to influence, right?

[00:08:18] [SPEAKER_01]: Yeah. And I think it's easy to say that. I think the other thing that goes along with that is what is it going to take for your organization to change?

[00:08:57] [SPEAKER_01]: Whether they have mainframe systems or client server on-premise systems where they can actually jump over a generation, move past just moving to the web and going straight to full cloud providers.

[00:09:09] [SPEAKER_01]: And then also the same on the ML and AI side, being able to leap past having some of the basic analytics and jumping past that.

[00:09:17] [SPEAKER_01]: But that takes a big devotion to be able to make those sort of jumps.

[00:09:21] [SPEAKER_01]: And it's, in many cases, it's less about the technology and more about changing people and their minds as well.

[00:09:27] [SPEAKER_00]: Totally. Well said.

[00:09:29] [SPEAKER_00]: And so really around exciting things, can you share some of the most exciting technological innovations you believe will significantly improve patient care and outcomes?

[00:09:39] [SPEAKER_01]: Well, unsurprisingly right now, because this has been such a hot topic, I think AI and ML do have a great deal of potential there.

[00:09:46] [SPEAKER_01]: I think one of the most interesting things, and maybe this is just a philosophical musing, is that for most of computer science's history, going back to what my dad did, data processing, which was called back in the day and punch cards and all of that sort of stuff and the paperless office.

[00:10:04] [SPEAKER_01]: The computers are based on a very specific premise, which is binary.

[00:10:09] [SPEAKER_01]: And this may sound overly technical, but the idea with binary is just on or off.

[00:10:15] [SPEAKER_01]: One or zero is just on or off.

[00:10:16] [SPEAKER_01]: It's just a switch, right?

[00:10:18] [SPEAKER_01]: And because of that, that has built on top of it both an expectation and a reality that computers should be infallible.

[00:10:28] [SPEAKER_01]: I'm not saying that's true or false or good or bad.

[00:10:31] [SPEAKER_01]: It's that if your machine gives you a different answer today than it did yesterday, that's a bug.

[00:10:39] [SPEAKER_01]: That's a problem, right?

[00:10:40] [SPEAKER_01]: Like that's an error.

[00:10:42] [SPEAKER_01]: Somebody coded it wrong.

[00:10:43] [SPEAKER_01]: Pentium bug.

[00:10:44] [SPEAKER_01]: The hardware had a problem.

[00:10:46] [SPEAKER_01]: You know, like it's a problem.

[00:10:47] [SPEAKER_01]: And now we look at AI, especially generative AI, and that's a feature.

[00:10:54] [SPEAKER_01]: Very much like humans, right?

[00:10:56] [SPEAKER_01]: Like people, you know, if you ask your significant other a question, the same question you asked yesterday, you may get a different answer today, right?

[00:11:05] [SPEAKER_01]: And the same if you're asked that question.

[00:11:07] [SPEAKER_01]: And that's considered, like that's fine.

[00:11:09] [SPEAKER_01]: There's emotions.

[00:11:09] [SPEAKER_01]: There's other things that go into it.

[00:11:11] [SPEAKER_01]: And it's actually a benefit.

[00:11:12] [SPEAKER_01]: We call that creativity in many cases.

[00:11:14] [SPEAKER_01]: It is the foundation of creativity.

[00:11:16] [SPEAKER_01]: And I think that that fundamental shift means that a foundation of what we believed about what computers are capable of and our expectation of them as humans is shifting too.

[00:11:31] [SPEAKER_01]: I think it's unfortunate that we think it's intelligence.

[00:11:35] [SPEAKER_01]: There's a great article written by Jaron Lanier, who was one of the forefathers of VR.

[00:11:40] [SPEAKER_01]: This was published about a year ago.

[00:11:41] [SPEAKER_01]: He said there is no such thing as AI.

[00:11:43] [SPEAKER_01]: He was bringing a thought process to what we think about AI.

[00:11:47] [SPEAKER_01]: AI is just human intelligence being able to run through an algorithm.

[00:11:53] [SPEAKER_01]: The intelligence is human.

[00:11:56] [SPEAKER_01]: He's proposing also that there's economic incentives for the AI companies and not for the humans doing the creating.

[00:12:04] [SPEAKER_01]: And long term, that has a very negative impact, which is there's less incentive for people to do podcasts, people to create art, people to...

[00:12:13] [SPEAKER_01]: And that actually is what's driving all the intelligence we see in AI.

[00:12:17] [SPEAKER_01]: And his argument was, if we stop calling it artificial intelligence and call it what it is, give it a different name where we're like, oh, no, that's based on human intelligence.

[00:12:27] [SPEAKER_01]: Then we can start to treat it in a different way and potentially change how we compensate and do all those sort of things.

[00:12:33] [SPEAKER_01]: Certainly goes into privacy issues.

[00:12:34] [SPEAKER_01]: And like many of the issues that are being faced in this hype cycle.

[00:12:37] [SPEAKER_01]: So I think that it is an amazing technology.

[00:12:40] [SPEAKER_01]: It's going to change so many fundamentals about how we use technology and how we think about it, just like mobile and internet did.

[00:12:46] [SPEAKER_01]: But we also have to apply the same caution like we do with social media and other things with young people to like we're opening Pandora's box.

[00:12:56] [SPEAKER_01]: So I'm excited about it, but also like wondering a little bit what the change is going to mean for us.

[00:13:02] [SPEAKER_00]: Yeah, no, definitely given us a lot to think about on this question for sure, Dave.

[00:13:06] [SPEAKER_00]: So thank you for your thoughtful response.

[00:13:08] [SPEAKER_00]: How should organizations address the critical issues of data security and privacy?

[00:13:14] [SPEAKER_00]: And do you have an example of what your organization is doing?

[00:13:18] [SPEAKER_01]: Yeah, so there's several levels to that.

[00:13:20] [SPEAKER_01]: So there's the individual privacy, which is a massive issue that's being heavily legislated.

[00:13:24] [SPEAKER_01]: And we're part of that legislation, too, because we have a particular perspective.

[00:13:28] [SPEAKER_01]: There's over 65 million users of the Workday system globally.

[00:13:32] [SPEAKER_01]: So there's a lot of human beings, workers in our case, that are part of the system.

[00:13:38] [SPEAKER_01]: So in general, legislatures, both here in the U.S. and overseas, trust us, like have a perspective.

[00:13:45] [SPEAKER_01]: Yeah.

[00:13:45] [SPEAKER_01]: You know, on, OK, well, how do you deal with this?

[00:13:47] [SPEAKER_01]: Because these large companies want you to take care of their data.

[00:13:50] [SPEAKER_01]: So we've been involved in California at a federal level as well in the legislation and, you know, heavily because we really do feel like there needs to be direction for it from an organization like ours.

[00:14:01] [SPEAKER_01]: We're not social media.

[00:14:02] [SPEAKER_01]: We're like, we're something different.

[00:14:04] [SPEAKER_01]: So there's that individual level.

[00:14:06] [SPEAKER_01]: But then there's also something that we care about deeply as well as the enterprise level, like the privacy and security and compliance of the data for large organizations.

[00:14:15] [SPEAKER_01]: And that's an area where we have a great deal of knowledge in addition to the individual as well.

[00:14:21] [SPEAKER_01]: And I think that this is an area, and I talked about, we have an agreement with our customers in terms of how we use that data.

[00:14:27] [SPEAKER_01]: And a lot of it has to do with transparency, a really deep understanding of what data we're using, what the purpose of that data is, regular reviews on whether we need that data anymore.

[00:14:40] [SPEAKER_01]: These algorithms are funny that way, right?

[00:14:42] [SPEAKER_01]: Like you start out, these particular fields are really critical for making decisions.

[00:14:47] [SPEAKER_01]: And then over time, you realize, you know what, for this company, they actually don't matter that much.

[00:14:51] [SPEAKER_01]: Well, why are you still collecting that data if you don't need to?

[00:14:55] [SPEAKER_01]: So we do regular reviews for the algorithms.

[00:14:57] [SPEAKER_01]: We also publish for every single one of the machine learning services that we have.

[00:15:03] [SPEAKER_01]: We publish exactly which pieces of data are being shared, why we justify that.

[00:15:10] [SPEAKER_01]: And then also say, here's the business benefit you gain out of it.

[00:15:13] [SPEAKER_01]: Because a lot of this is discussions with chief security officer, chief information security officer, chief privacy officer.

[00:15:20] [SPEAKER_01]: They need to know.

[00:15:21] [SPEAKER_01]: They need to know, well, what am I sharing?

[00:15:23] [SPEAKER_01]: And then, okay, I'm taking some risk by sharing data.

[00:15:26] [SPEAKER_01]: What's the benefit of this?

[00:15:28] [SPEAKER_01]: And those sort of practices, I think, help organizations hold us as a vendor accountable.

[00:15:34] [SPEAKER_01]: Because the other benefit that Workday has is we only have one way of providing our service, as I said, right?

[00:15:41] [SPEAKER_01]: It's in the cloud.

[00:15:42] [SPEAKER_01]: Everybody has the same data model.

[00:15:43] [SPEAKER_01]: Everybody has the same structure.

[00:15:45] [SPEAKER_01]: It's all consistent.

[00:15:47] [SPEAKER_01]: Everybody's running on the same version.

[00:15:48] [SPEAKER_01]: And that means that whatever crazy requirements some of the largest companies in the world,

[00:15:54] [SPEAKER_01]: banks and pharmaceutical companies and payers and providers, whatever they have, it applies for every single one of our customers as well.

[00:16:02] [SPEAKER_01]: And so we have to be really thoughtful in terms of how we apply those things.

[00:16:06] [SPEAKER_01]: And it needs to apply globally, too.

[00:16:08] [SPEAKER_00]: Great examples.

[00:16:09] [SPEAKER_00]: And really just highlighting the benefit of standardization.

[00:16:13] [SPEAKER_00]: You guys decided you wanted to be on the cloud.

[00:16:15] [SPEAKER_00]: It was a strategy.

[00:16:16] [SPEAKER_00]: Going back to how do we use AI and ML?

[00:16:19] [SPEAKER_00]: And at the end of the day, human intelligence, I think, will win the day.

[00:16:23] [SPEAKER_00]: Dave, if you wanted to just leave us with a closing thought, what would that be?

[00:16:29] [SPEAKER_00]: And then where's the best place that the listeners could get in touch with you and the company to learn more?

[00:16:34] [SPEAKER_01]: Yeah.

[00:16:34] [SPEAKER_01]: So the last thought I'll leave you with is one of our guiding principles at Workday, which is leaving a human in the loop when it comes to AI and ML.

[00:16:41] [SPEAKER_01]: And my personal view on it is so far, AI doesn't seem to have judgment.

[00:16:47] [SPEAKER_01]: It has many other things.

[00:16:49] [SPEAKER_01]: But judgment is uniquely today a human trait.

[00:16:52] [SPEAKER_01]: And we want to make sure that that remains a critical part.

[00:16:55] [SPEAKER_01]: And it's really surfacing the information for a human to apply that judgment and be able to make decisions.

[00:17:01] [SPEAKER_01]: And you can learn more at Workday.com about our solutions, finance, HR, payroll for enterprises, all delivered through the cloud.

[00:17:10] [SPEAKER_01]: Amazing, Dave.

[00:17:11] [SPEAKER_01]: I want to thank you for spending time with us today.

[00:17:13] [SPEAKER_01]: Thank you so much.

[00:17:14] [SPEAKER_01]: I really appreciate it, Saul.

[00:17:15] Thank you.