Workday Podcast: Putting Customers First in AI Transformation

Jim Stratton: As Workday continues to advance AI innovation, we recognize the importance of transparency, explainability, and control. Ensuring these elements are integrated into our AI solutions is crucial to meeting our customers' needs. Today, we're here to discuss how we prioritize these aspects during our customer's AI transformation journey. I'm Jim Stratton, Chief Technology Officer at Workday. Joining me on the Workday podcast is Andy Cannon, our Vice President and Deputy General Counsel of Product and Technology. Together, Andy and I will explore how we put our customers first in our AI initiatives. We'll share insights into Workday's approach to AI, including our commitment to responsible implementation, data protection, and transparency. So let's jump right in. Hi, Andy, welcome. Thanks for joining today

Andy Cannon: Thanks for having me, Jim.

Stratton: So let's talk about the AI transformation that's taking place right now. We've talked to so many organizations about AI. What are we hearing at the big-picture level?

Cannon: So I think, from my perspective, it's all about customer expectations. So the first thing is I think we've noticed by talking to all of these customers over the years is that different organizations are at different stages of their AI journey. Some organizations are ready to increase the productivity of their end users right away and their employee base, and others are a little bit-- they're not quite there yet, and we definitely understand that. And so we want to be ready when the customers are ready. The customers that are ready right now, we're ready to give them the AI features right away, and the customers that are not quite ready, we want to give them the tools so that they are ready and that they can increase the productivity in their workforce, just like the customers that are adopting it right away.

Stratton: So let's talk a little bit about responsibly using AI. For many of our customers, data governance is really the first step in how to responsibly use AI. But what exactly does that mean, and how can customers get comfortable with it?

Cannon: Well, as you know, I mean, you and I talk about this all the time. Data governance is really, really near and dear to my heart, and I think that, when you talk about responsible AI, data is key. So our customers need to understand not only the AI features that they might uptake for their organization, and I think that goes to the explainability aspect. We have AI fact sheets out there that our customers can look at. They can decide what is right for their organization. But also, we want them to understand the data that we're using from their organization to personalize these AI features for them. So we know what we see in the industry-- and I think I have a little bit of a unique perspective because even at Workday, I get to see what our competitors do. And not everybody is transparent about the data that's being used out there, and so I think that that's something that we've thought long and hard about. And we want to give customers visibility into their data. We want to give them visibility into the AI features and how they're used and how they can support their workforce to increase productivity across the board.

Stratton: Yeah, it's a great point, and it's really key to how we maintain and continue to develop trust with our customers in the solutions that we're building. So along those lines, Workday has worked really closely in collaboration with our customers to improve our AI offerings and to improve the overall customer experience. How's this collaboration evolved to further empower customers in controlling their data and creating a robust data governance plan while also optimizing the benefits they get out of the service?

Cannon: So I think, at Workday, with our customers, trust is key, so we want to make sure that we keep the same level of trust with all the offerings that we deliver to our customers. And so, as I was talking about this data governance aspect, when we talk to customers, we encourage them to put together a governance plan, just like we have at Workday, where you can get your security professionals, your compliance professionals, your legal professionals, your privacy professionals involved in the process so that there is no barrier to adoption. And I think that that is one of the main things that we see out there, especially even when Workday is uptaking AI features internally, is that if you don't have a governance plan in place and you don't have the materials from your vendors, you are going to have a barrier to adoption. And if you cannot get these other organizations or these other cross-functional groups in your organizations the materials that they need to evaluate these offerings, then you're just not going to be able to uptake it. And I think that that's a challenge that everyone has when you're trying to increase the productivity of your workforce using all the AI that we hear about in the industry today.

So one of the things that I'm pleased to announce - and I know you've been talking about it quite a bit with our customers - is that we have a new universal MSA contracting structure, and this contracting structure is basically-- it's a way for customers to leverage AI with every single feature that they use in Workday. So we heard Carl in our earnings report talk about our 50 generally available AI features right now across the board with all of our product lines and then another 25 generative AI capabilities that are coming very, very shortly. So the catalog is growing and growing. And now with the universal MSA, the customers will be able to sign a contract with Workday; the AI will be ready for them when they are ready.

And then the thing that I'm actually most excited about right now is the data governance UI that your team put together. So it was a great collaboration with legal and compliance, security and privacy, but what we've built out for the customers now is basically a one-stop shop where you can see the AI features that are available to you in a user interface right away. So customers can look at these AI features, they can decide what's right for their organization depending on their implementation, and then, most importantly, every single AI feature reports on the data that we're leveraging to personalize it for them. And I think that that's really, really powerful. We have reports that can be downloaded, delivered to the compliance functions and security and privacy in their org. And then, even better, when things change, what we've seen with some of our competitors out there is that you have to subscribe to a certain website, you have to look at the same information over and over and decide what's changed. But that's not what we're doing at Workday.

We want to meet the customers where they are, and so whenever there's a change to the data, whenever there's a change to a feature, whenever there's a new feature, customers can set up a security group. Their chief privacy officer, maybe their head of legal, maybe even their CTO - who knows - whoever wants to be on this list will receive an in tenant notification. And then if you have your implementation set up this way, you'll even get an email. So when things change, we let the customers know. We give them time to decide and make a decision if they want to continue to leverage this AI and how they want the AI features personalized for them. So I think it's really, really powerful. It's something that I really haven't seen from our competitors or other organizations in the industry. And I think it really will add a huge benefit to our customers where you can set up a governance plan natively in Workday and get your organizations comfortable to what it is we're delivering.

Stratton: Yeah, that governance built into the platform and the product that they're already using is hugely powerful for customers. That's great. So you've mentioned transparency and trust a couple of times. They're obviously both pivotal in developing and deploying responsible AI. So can you talk a little bit about how Workday ensures customers' understanding and confidence in how their data is being used, kind of further to what you were just talking about, so they understand how the data is used for AI algorithms without compromising privacy or ethical standards?

Cannon: Well, I said it before, but I think trust is a very important element to this. And so, along with the universal MSA, along with the data governance UI where customers can natively see in Workday what it is that we're doing with the data to personalize for them, on top of that, we have a whole bunch of collateral that we're developing and that we have developed for customers. And so I touched on it, but we have AI fact sheets for every single AI feature that we're delivering to our customers. It goes to the explainability aspect of it where customers can see what it is we're building and how it can be used in their organization. But on top of that, we talk about the data that we've used to train the models. We talk about the privacy-enhancing technologies that we use. We talk about security. We talk about responsible AI. So the point of it is that you can go on our customer community website, you can download these PDFs, you can deliver it to these other cross-functional groups in your organization, and you can really get them comfortable quickly. And I think, for us, we want our customers to continue to trust Workday, and we want them to understand that, just like any other feature at Workday that they're already leveraging, they're still going to get the same world-class privacy, they're going to still get the same world-class security, and they're going to still get the same world-class responsible AI program that you hear about all the time when you hear about Workday.

Stratton: Yeah, that's great. So now I want to circle back to one of the other points you mentioned. So organizations are in different stages on their own journey to using AI, so let's talk about that and dive into that a little bit more. How do we meet the needs of customers along that spectrum of where they are? For example, how do we make sure that we're serving the needs of the customer who wants AI now and is ready to run today while also meeting the needs of the customer who's not quite ready yet and have it available for them when they are ready to jump in?

Cannon: So I think the first step of the journey is just delivering a contractual structure where customers can sign up for an AI-enabled Workday. In the past, customers would sign up for Workday, and then they would have to sign it in a separate contract with our AI, and it was kind of an add-on. And I think what we've heard from our customers and what we've seen when we go to Rising, when we go to other events, is that customers want to leverage AI right away, and they don't want to go through a separate process. And so I think that the first piece that's very, very powerful, is that once customers sign up with Workday, they have the expectation of an AI-enabled Workday, and we're delivering it to them.

Cannon: So on top of that, I think the data governance UI that we were talking about is extremely powerful in that customers have complete control over their data. So just like any other Workday feature, the customers can enable the features, and then they can decide on the data they would like us to use to personalize it. So this aspect of control, I think, is something that we've always based AI at Workday on, and that continues under this new program. For the customers that are ready, it's going to be ready right away, and customers can enable these features and start leveraging the benefits of enhancing their workforce.

Cannon: For the customers that are not quite ready, that have a governance plan in place, that are probably looking at our features and deciding whether it's right for them, they can simply go into the UI and just turn off the data contribution. They can decide not to enable the features. And like we were talking about before, it's ready when they are. So I think what we hope is that, by giving customers all of this collateral right at their fingertips, by giving them a UI that they can leverage with their staff right away, when they are ready, they can turn them on. And maybe it's one or two things just to start with. Maybe it's the full suite. But we don't want to have any other barrier to adoption because we know that AI is coming quickly and customers are going to want to leverage it.

Stratton: And it really puts the control directly in the hands of the customer.

Cannon: Yes.

Stratton: That's great. So when we think about helping customers on that AI readiness journey, one of the big factors on customers' minds is generative AI, gen AI. We've talked about the hype cycle around gen AI before and how Workday is really focused on how we can deliver value to our customers. But how does what we've evolved here in terms of the new UI and universal MSA help customers from a gen AI perspective?

Cannon: So as I've been looking at all the generative AI features that your team's building, Jim, something kind of hit me, is that we are delivering these features to our customers natively in Workday. And whether our customers know it or not, their workforce is using generative AI right now. And it's kind of been a Wild, Wild West out there where these startups are popping up, and things are coming out every day. And even I'm at home with my 10-year-old and 8-year-old, and they say, "Hey, Dad, look at ChatGPT and how cool it is." And so everybody's using it right now, and I think it's really increasing the productivity of everyone. And as a legal professional with a legal and compliance team, we have to review these features to be able to be leveraged at Workday. And that's what our customers have to do, too. And I think it's a really daunting task where you have to go out there and look at these vendors, and just like our customers and kind of their readiness journey, some of these vendors are in different places of their readiness journey, too.

And so not everyone has the same trust story that Workday does, not everyone has the same world-class privacy and security, and so whether they like it or not, their end users are using it already. And so when I see stuff like our generative AI features that are coming out, like content generation, summarization, chatbots, things like that, that are really going to improve the experience in Workday, why not just do it in Workday? It's the same security. It's the same privacy that our customers have known to grow and love. It's the same trust story. We're treating it like all of our other products. And if I was in charge of an organization and deciding what AI features we were using, I would love to be able to leverage our trusted vendors rather than have to go out and look at vendors that may not be at the same readiness level that someone like Workday is.

Stratton: I mean, Workday really was founded on the principles of keeping customers' data secure and private from day zero of the company, and that doesn't change in the world of AI, and you mentioned that a few times. How do we really uphold that commitment for training Workday AI across selected product lines or individual features? How do we maintain that in the world of AI?

Cannon: Well, I think about some of the sessions that you did at Rising when you talked about what our plan is and how we're going to be delivering AI. And I talk with folks on your team all the time. And I think the one thing that I want to make clear to people that sit in my seat at some of these other organizations is that Workday is not delivering large foundational models like a ChatGPT that can do a myriad of things. When you have something like that, you have issues with hallucinations. You have issues with, potentially, security of data and things like that. But that's not what we're doing. We're building based on use cases. We're using best-in-class methods like prompt engineering, retrieval-augmented generation. And Workday is in control of not only the inputs that our customers put in but also the outputs. So we're ensuring that-- before we deliver these features to any of our customers, we're ensuring that we're getting it right. We're ensuring that it's safe. We're ensuring that there's no data leakage anywhere. Like you said, we're upholding the same level of security and privacy that we do across the board. But honestly, this feels a little bit weird for me telling you what we're doing because you're kind of the expert on this and I listen to it. So maybe I'll throw it back to you. I think it's a really powerful story that we're not just taking something out of the box and delivering it to our customers and just hoping it works.

Stratton: Yeah. You're absolutely right. And the point that you made about these very large language models, we're not trying to build general-purpose solvers. We don't need to. And there is certainly a lot of value and a place for that. But within what we're trying to solve for, we're trying to solve for specific business use cases that drive direct value into our customers. And so that allows us to be much more constrained about what data we use to train those models and, as you mentioned, how we allow users to interact with those models. So very tightly constraining the prompts that go into interacting with those models. We maintain very tight security and privacy controls over the data flows through our training environments and out to the end inference servicing of those applications as well. So like I was saying before, it really comes back to how we built the company around protecting and maintaining security, privacy of this incredibly important business data and individual data that we host on behalf of our customers. And that flows through everything that we do, how we develop products, how we test them, how we deploy them out to production, how we communicate about what we're doing. It's just baked into who we are as a company and how we build products. And that doesn't change in the world of AI.

Cannon: Yeah. I mean, that's just really powerful for me. And I think, from a legal and compliance perspective, it just makes my job that much easier.

Stratton: Well, we appreciate the partnership. It's been a great conversation. Thanks for joining me again today, Andy. And thanks to all of you who are listening. If you liked what you heard today, be sure to follow Workday wherever you listen to your favorite podcasts. And remember, you can find more talks at workday.com/podcasts. I hope you all have a great work day.