Video: Workday CEO Aneel Bhusri on Machine Learning and Corporate Culture (CxOTalk #359)
We are speaking about the impact of machine learning on business. Aneel Bhusri, our guest today, is one of the most successful entrepreneurs in the world. Aneel is the cofounder and the CEO of Workday. Workday was started back in 2005 around the idea of bringing enterprise applications to the cloud. In particular, HR and finance, and not the point systems but actually the core systems of record around HR and finance. In particular, we focused in on the needs of the largest companies in the world because my cofounder Dave Duffield and I had a background at PeopleSoft where that's the market we knew.
While it seems pretty straightforward today, back in 2005, it was not so straightforward to think about serving the largest companies in the world with this new cloud model. We build the company around a core set of values that remain intact today, starting with employees being number one. Our view simple view is, happy employees lead to happy customers. Customers are obviously very important to us. Integrity, we try to do business the right way. Fun, we like to have fun. We tell people, "If you don't want to have fun, go work for a competitor." We like to innovate. That's what Workday stands for. Lastly, profitability. That's really necessarily the biggest core value or the most important, but it pays for everything else. We set about to do that around 2005 around those core values and around that strategy. The strategy has expanded to include planning now and analytics and spend management, but that core of HR and finance in the cloud is still really what got us going. After you started the company and just getting things going, I remember thinking at the time, "What an incredibly audacious goal it was at that time to build the complexity of financials and HCM at the scale that you do for these very large customers in the cloud." I guess we didn't think about it as audacious.
Maybe, in hindsight, it looked that way, but entrepreneurs are optimistic and Dave and I are optimists. What we saw was this trend that the legacy systems were increasingly breaking down for our customers while we were at PeopleSoft. Customers were having a hard time with upgrades. The user experience wasn't great. At the time, there was this emerging company, Salesforce.com, which is obviously now a giant, pioneering this new model. We thought, "Hey, that's the model of the future and someone needs to apply it to ERP. Why not us?" Aneel, at Workday Rising, one of the key themes that you spoke about was machine learning. Why is that such an important topic for you? I think it's the next frontier for enterprise applications and, at the end of the day, we're in business to help our customers run their businesses and their organizations better and smarter. If the last 15 years of the cloud have been about moving business processes from on-premise to the cloud and making it easier, friendlier, and better analytics, I think the next 15 years is about using your data to make better business decisions.
Machine learning is really the cornerstone set of technologies that help you understand your data, find insight in your data, and help you make better business decision that will hopefully lead to better business results. Is the key then focusing on taking the data and figuring out how you can make better decisions? Is that ultimately the point? There's a great book written by three professors from the University of Toronto. One of our speakers last week was Professor Ajay Agrawal from the University of Toronto. They wrote this great book Prediction Machines and I think that book spells out really well the power of machine learning. I don't think machine learning is about replacing humans. I really believe that it's more about a very positive relationship between machines and humans. Machines are really good at sifting through massive amounts of data and looking for patterns and looking for insight. Then humans apply judgment, which machines don't have.
You have a better prediction applied with human judgment should lead to a better outcome. I think it's kind of that straightforward. Really, what it's about for a company is: • Organizing all their data and looking for insights that could be predicting what might happen from a business event. • It could be anomaly detection. • It could be intelligent automation. Three different categories of machine learning that will all make your business better or your organization's decision making better. One of the areas where we've used machine learning is all around career succession planning where the machine can look through massive amounts of data because of all the computing power that's available today, look at people that have had successful careers at the company and, for you as the employee, saying, "Based on all that data that we've looked at, that would take a human probably years, if not decades, to sort through all that data.
They can do it in seconds. The machines can do it in seconds; predict the next logical step for you in your career and then make that recommendation to you. As a human, you can say, "Yeah, I want to take that next step with my career. I want to do something different." That's the power of machine learning where it sifts through massive amounts of data in a short amount of time to make a prediction that you can take action on that really is not possible without the computing power that's available today. It wasn't even possible just four or five years ago. Aneel, you've just boiled down this complexity to what is essentially a software feature, a very valuable software feature. Press a button and it gives a result. As somebody who is developing these products, can you kind of expose a little bit for us what goes on behind the scenes in order to enable this relatively simple concept that you've just described? It's a lot of complicated software technology but, at the highest level, it's about algorithms that are learning, that you're not programming them for a specific outcome.
You're having them learn from the data and, over time, you're training the algorithms to make better and better predictions. The more data you have, the more you train that algorithm, so if the algorithm is predicting the next steps in a career for an individual, if you've got X amount of data, you'll get a certain prediction. If you've got 10X, I can guarantee you you're going to get a better prediction. Really, it is about a series of algorithms. They're available on the Internet today through Google, Amazon, and Microsoft. The one that's really well known is Google TensorFlow where you could use these algorithms. Apply data against it. The algorithms will train itself to make better and better insight.
The value then is not so much, or the differentiation is not so much, in the algorithms but, rather, in the data that's being applied. The data and also knowing how to interpret the results. In our case, the data is around HR and financial data. Being able to know what questions that you might want to have answered and to know what the results might look like, that's really powerful. For us, we're very focused on the machine learning solutions to be basically application and context-aware. What does that mean, application and context-aware, at Workday? It knows the domain. We know the domain to know what questions we're looking to possibly get insight. The algorithms themselves are fairly standardized across multiple disciplines, but we have to apply our knowledge of HR and finance to know, "Hey, we're looking for something around career succession. We're looking at something around retention.
We're looking for something around faster audits. We're looking for anomaly detection on the posting process and auditing process for your financial systems." All those kinds of things, we know what we're looking for and so we're training the algorithm to give us insight into a specific business process or business outcome that we're looking for. Aneel, how do you decide where to apply machine learning? Your software covers so much ground. How do you make those decisions? It's at the intersection of where we can create business value for our customers. There are a lot of predictions that probably can be made, but they might not be valuable, so it has to be something that's valuable from a business perspective coupled with an ethical framework to make sure that we're doing it in a way that is ethical in the way that the data is being used. We recognize it's not our data. It's the customer's data, so we have a joint understanding of what the ethics are around the use of data from the customers. Maybe it's personal data of their employees. It might be financial data.
That ethical context is really important in the world of machine learning and artificial intelligence. Obviously, the confidentiality of the data is pretty clear. What is it about machine learning that creates this set of ethical issues that you feel are very important to address and manage? Say we're making predictions around people in their roles, whether it's career or whether it's retention. First of all, people should know their data is being used that way. I think transparency is a big part of the right ethical framework. Then I think the second piece is opt-in. If an employee or someone who is being tracked in the system doesn't want to have their data being used for machine learning purposes, they should be able to opt-out. Actually, forcing an opt-in is even more powerful than an opt-out. Opt-out, you have to go to work to opt-out. In an opt-in, you proactively say, "I want my data to be used for these machine learning purposes because I trust what the organization is doing with that data. It's all built around trust and transparency. From our view, it has to be opt-in and you have to be transparent with the employees how their data is being used. This issue of trust is really important.
Can you contrast at all the nature of trust when you're talking about machine learning and this type of data versus trust and confidence in you as a company outside the machine learning domain? What happens that's unique? I think they go hand-in-hand. I think our customers trust us to be good stewards of their data and how it transcends into, as they do their analysis on their data, we're going to continue that focus on integrity, that focus on ethical framework, and we're, frankly, also going to help them sort it through because everybody is doing machine learning on behalf of their companies. I think every company in the world should be using machine learning at some level or they will be left in the dust at some level. As part of this endeavor around machine learning, they've got to develop an ethical framework, and so I'd recommend every company has a chief ethics officer that helps them sort through these challenging questions.
I think it transcends the way we've worked historically with the customers around machine learning is really going to be no different but I think we're jointly going to have to discover the right way to deal with it in a really ethical way because it's really a new domain for everybody. Is that because of the quantity and the type of data that you're now aggregating that, in the past, you didn't? We've always aggregated it and we've always considered it our customer's data but it was much more around automation, making a business process more efficient, producing better analytics. It wasn't about making predictions on that data. Now, I think we're in a place with personal data and, in particular, where we have to be very careful. As we've seen on the consumer Internet, they haven't all been perfect with the data protection and privacy issues of their consumers. I think you see how they're paying the price. We're not going to get anywhere close to that line.
We've got to just do it the right way. Yeah, I think the biggest change is, we've aggregated data in the past but not for this purpose. You're thinking very strategically and carefully about precisely where you draw those lines. Yes, exactly, and we have both really strong experts in privacy and data security and now we've also got a chief ethics officer that helps us make sure that, across the company, we're doing it in the right way. There's this whole new layer of this trust and transparency. Recognizing that we are a data steward, right? The data that our customers store on their Workday systems, it's their data. We're not looking to make money from selling the data. That's not the business we're in. We're looking to help our customers gain value out of their data and gain insight out of their data, so it's much more of a partnership and, in some ways, a research project to say, "What are you looking for to make your business better? Let's go see if we can use machine learning to help you get some insight into how your business has been performing and where you want it to go.
Maybe we could be helpful with some of these machine learning algorithms to help make some of those predictions." You said a few minutes ago that, in the future, every company must embrace machine learning. I don't think you said, "To survive," but essentially to thrive in the future. I think so. If a company is using machine learning as an example for intelligent automation or speeding up the process of a financial audit and the other company is not, that other company is going to be disadvantaged. If one company is using machine learning to get a better handle on the talent they have, maybe they're looking at the inventory of skills they have in the company and looking for the shortages that they might have as a company in 12 months or 18 months and another company is not, well, the one company that is doing it is going to have a huge advantage. Case-by-case, if you're using it, it's going to lead to better business outcomes because, again, these machine learning algorithms are all about predictions and humans applying judgment against the predictions.
If the predictions are better, the judgment is more valuable. It'll lead to a better business decision, which will lead to a better business result. I think it's really that straightforward. If you're not using these technologies, you're going to be flying blind. If you look at the history of analytics, while people have talked about predictive analytics over the last 20 years, as you know, most of it was really trying to look at the past and use that as a way to guess at the future. The new kind of machine learning algorithms really will give you a view into what is going to happen with your business with a high degree of confidence if the data sets are right. It really is predictive. If you have that predictive engine and your competitor doesn't, I'd bet on the one that does have it. I'm assuming, and correct me if I'm wrong, you're going through the Workday suite and trying to identify where those points that you can apply machine learning based on the criteria, as you said, of what will provide the most value for your customer.
Are you doing that, dissecting in that way? Absolutely, and we're testing it out with our customers. We're making sure that, when we come up with a use case, we then check with our customers who are really using machine learning and say, "Would this use case be valuable for you?" There is quite a bit of iteration with the machine learning discipline. You want to figure out if you have the right data and are you solving the right problem. You're constantly looking for that combination. It's the right data. You need a lot of data. Then you have the right business problems to solve. You get those three things together; it can be really powerful but it isn't always obvious how to get the right three things together. The only way to do it is with a lot of customer interaction. What about confidence in the result, because I don't have any way of knowing how you arrived at that answer and can I trust and have confidence in that result? That's part of being transparent about how we're running the algorithms and what data sets we're looking at. Also, I take the analogy of weather.
We do trust weather reports now. If you look over the last 20 years, they've become increasingly more accurate. Today, The Weather Channel will tell you, "It's going to start raining at 8:07 a.m.," and it starts raining right at 8:07 a.m. Twenty years ago, it was a guess. Today, these technologies, because they're so good, are very predictive and those are machine learning technologies that are predicting weather. I think, as people see the results actually pan out, they'll gain more confidence in the algorithms. I also think that, in all the predictions, there'll be a confidence level associated with it. It might be, we're 90% confident. With 90% confidence, we think this is happening in your business. It's always hard to get to 100% confidence level. Even when you look at the weather forecast, they'll tell you 50% chance of rain at 10 o'clock and 90% chance. Very few times they'll say 100% but, when it is 100%, it's almost is always raining. I do think there's a level of, you're going to have to look at the result and you have to look at the level of confidence around that result too.
That is part of the machine learning process. You're targeting or, I should say, your customers are very often chief financial officers and chief human resource officers. Where are they, would you say, just in general, in terms of adopting these types of technologies and being comfortable with it? Everybody is becoming a technology company. You can't just say you're going to leave technology off to the side. Every company, even if they're in retail, they're becoming a technology company. If they're in distribution, they're becoming a technology company. Technology is impacting everybody. I'd say, first and foremost, that trend is happening across all businesses, all organizations. That's leading to, actually, frankly, to finance organizations and HR organizations that are much more data-aware and data-centric than maybe they were 20 years ago. In almost all the cases when you go in and talk to a potential customer or a customer about using machine learning, they're already doing it somewhere in their business.
They might not be doing it yet in HR and finance, but they're doing it somewhere in their business. • If they're an insurance company, they're definitely using machine learning in their claims processing world to improve on the actuarial tables that they've had for decades. • If you're in the oil and gas world, you're using it to do better predictions on where you might find natural gas. If you're in the media world like a Netflix, they've been using machine learning to predict what show you're going to want to watch next for the last decade. It's already being used somewhere within the organization. What you find is, IT is already on top of the technology and HR and finance are ready and they have proof points about how it's valuable already, usually to the core part of the business, and now we're bringing that technology to HR, finance and spend, and planning and analytics, all the areas that Workday is involved with. There is a growing understanding of the role of machine learning and especially the role of data.
Yes. There is probably a CTO, a CIO, or a chief data officer somewhere in almost every company that's already down that path and it's extending it to the worlds that Workday is focused on. I would say what I've been pleasantly surprised by is how much on top of this new world the CEOs are. They understand the power of data and, in particular, to whatever their core line of business is and they're trying to use machine learning there. When you have the conversation with them about machine learning for HR and finance and spend, they totally get it. Is it forcing these CEOs to become technologists? I don't know about technologists, but technology literate and knowing what they need on their teams in terms of technology expertise. Why do we have so much fear out there that, when it comes to machine learning, the robots are going to take our jobs? As you know, we've got a big investment in this program at Workday called opportunity onramps where we take veterans.
We take caregivers who have been out of the workforce for a long time. We take young adults who might be underserved or have not had the opportunities that they should have, and we try to train them on the modern technologies so they can find roles that are part of the tech world rather than not. Does machine learning make things easier for your customers or create complexity? It definitely creates complexity but, I think, at the end, we're all trying to be better in our organizations. We're trying to be smarter. We're trying to be more efficient. Machine learning definitely pushes us further. Cloud got us to a place where our business processes can now keep up with this everchanging world where legacy systems got stuck in mud. Cloud, we come up with an update every six months. We're keeping up with all the modern business requirements, all the new legislation, all the changes that are happening. The cloud gets you there.
The cloud also captures all your data in a way that's really usable to get at this insight. Now, you have all this data. Why not try to gain insight into it? That's why I think this machine learning movement is really the biggest thing that's happened since the cloud. It might make our customers' lives more complex to deal with the data but, in the end, I think it makes it easier to make good business decisions. We're all trying to make business decisions based on the best analysis that we can get and the best data that we have. Machine learning gives you the best analysis and the best data. That's what it's for there. Obviously, it creates complexity for you as a software developer but you're encapsulating it so that it's presumably really simple for the customers. We're actually trying to simplify it for our customers, but it is a new discipline, right? If the last 20 years were about coming out of college with a computer science degree because that was the place to be, the world still needs a lot more computer scientists but we now need a lot of applied math majors who can turn into data scientists.
It's a new muscle for all of our customers to not just be about automating their business processes but actually getting value out of their data. This whole decision science world is exploding. If I were coming out of college today, I don't know. I came out with an engineering background. I might be more focused on applied math. We have some questions from Twitter. Zachary Jeans is asking, "Did you learn or observe anything particular at Workday Rising that inspired you about the future?" I'd say a couple of things. Talking to our customers is always inspiring because they give you great insight and hope that what we're working on really does matter. It's making their lives better.
They also give us input on how we could do better, in general. Hopefully, you felt our customers were happy with 97% customer satisfaction. That's really meaningful for us. Of all the speakers we had, I found Lin-Manuel Miranda to be incredibly inspirational, his message of inclusion, his message of creativity. He said something that I thought was really powerful. He was asked by the interviewer, Soledad O'Brien, "Hwy is it that so much of your work is timeless?" He said, "I never think about trying to create timeless work. I focus on what I love and I try to do it with integrity and honesty. I believe if you do it that way, what you work on has a chance to be timeless." I found that very inspirational. I found him to be very inspirational. We have a few other questions from Twitter.
When companies are looking for opportunities with machine learning, how do you avoid doing a machine learning project just for the sake of machine learning that does not necessarily deliver the business value? You walk through with the executives what problems you're trying to solve. I do think it comes down to, let's take a step back and figure out what problem are you trying to solve. Are you trying to solve a product quality issue? Are you trying to solve a geographic revenue issue? Are you trying to solve a talent issue? Then you work your way back to the data that you might have and how we might use machine learning to solve it. You always have to start with, what business problems are you trying to solve? From that standpoint, it's no different than applying any other technology, being clear about what you're trying to do. Exactly. Exactly. In some ways, machine learning is this new shiny object and, with any new technology, there's always the risk of getting caught up with that rather than the business value.
Exactly, although, I do think there is value in experimenting in different areas because it is such a new area. You might actually find value in an area that might not make intuitive sense, so there is a value in experimentation and iteration as long as you stand by the ethics rules and don't get into a gray area there because I think the ethics area is actually pretty black and white. What is the black and white? Can you define that? I always think about the Wall Street Journal test. If it was written up in the Wall Street Journal that you're using data this way, how would people react? I think that's usually a pretty good test. That's why, if there's a line between right and wrong, we don't want to get anywhere near that line. We want to stay clearly on the right side of using artificial intelligence in a good way. As you know, Michael, I've always talked about technology as being neutral to good or bad, right? You can use technology for good purposes. You can use it for bad purposes.
You could just see that with the security hackers versus the great companies that are building great security products. You have people using similar technologies on both sides of the spectrum. I think it's important, as a vendor, that we are just clearly on the side of using technology for good. Well, that's a very clear statement. We have another question from Twitter. Arsalan Khan asks, "Do you think the role of enterprise architects diminishes or is enhanced by using machine learning and AI to gain operational excellence?" Absolutely enhanced. A big part of, to me, what an enterprise architect needs to do is understand the enterprise, understand all the data sources, and figure out where machine learning can be applied. I think it's even more critical in a world of disparate systems for someone to have an overarching view of how all these systems come together and, frankly, how all the data comes together.
Let's talk about employee relationships. It goes back to our founding principles and I have to give my cofounder Dave Duffield a lot of credit. He pioneered this employee-centric, employee-first culture at PeopleSoft and we really brought it forward to Workday. The basic premise is that you can talk about customers being number but, if your employees are unhappy, it's really hard for unhappy people to make customers happy. If you start with happy employees, you empower them, you give them all the right tools, you motivate them, you educate them along the way in new disciplines, it's amazing what they can do. Our simple view is, happy employees lead to happy customers and it's paid off. We're number four this past year on a great place to work, the survey in Fortune Magazine as a great place to work for employees. What I'm proud of is, we're a great place to work across lots of categories: women, millennials, diverse backgrounds, across working parents.
We're really a company built around inclusion and diversity. You take those empowered and happy employees and you say, "Your number one job is making customers happy." Well, that very happy base of employees led to us having a 97% customer satisfaction level. Maybe there are other ways to do it, but this way definitely works for us and it's worked for us across two companies. It also leads to a lot of business benefits too. Our turnover is much lower than our major competitors. Our ability to recruit is higher because we're viewed as a great place to work. There are just lots of benefits as a result of being a great place to work. How do you do it? Is there a formula? Is there a specific set of steps? What's the secret to doing that? The first thing is, when you hire people, make sure that they fit your value system.
I think you know the first 500 people that were hired at Workday, Dave and I interviewed all of them. We didn't interview them for their skills. We assumed the managers got that right. We interviewed them to make sure that they fit our values. That's where you start. For managers, in particular, we have intensive manager training. Once a year summits, the people leadership summit where we walk through how you manage at Workday, how you manage with compassion, how you manage by being a good listener and being empathetic, how you deal with conflict. All the ways that are unique to our culture and our value system, we try to teach managers how to do that. Then we have an ongoing program called Ignition where you're learning along the way about not just being better at whatever your discipline is, marketing or engineering, but also how do you become a better manager and leader. My simple commitment to our individual contributors who do the bulk of the real work, they deserve a good manager.
That's a big part of how we make it all work is we invest heavily in manager training and making sure we're hiring the right people. Have you experienced bumps in the road or, as you observe other companies, are there common challenges or what are the common challenges that companies tend to face when they have this as an aspiration but they try to execute on it? It's challenging during rapid growth. We're now almost 12,000 employees. A few years ago, we were just a couple of thousand. We've on-boarded a lot of employees the last few years. We've definitely had bumps along the way. One of the bumps was a drop in some of our rankings. As we dug into it, we recognized that while, in the previous two years, half the employees were new, half the managers were also new. If they were coming from companies outside of Workday, they didn't always manage in a Workday way with the values and focus that we do. That was the impetus for creating the People Leadership Summit that our HR leaders, Ashley Goldsmith and Greg Pryor, really pioneered and it's worked out really well.
You've just got to pay attention. Back to machine learning; you've got to pay attention to the data and the data was saying that we were drifting on the experience of our employees. We're always looking at that data. We do a pulse survey using the Workday system. Every week, we ask two different questions of our employees to get a pulse of how they're feeling about the work and it drives employee engagement. It drives the data to help us make sure we're doing all the right things. The leadership of the company has to care about it and be willing to invest. That's pretty unusual. People pay lip service to it but, in practice, I don't see it all that much. I think it might have been unusual and I give Dave all the credit in the world. He pioneered this way back when it wasn't fashionable to be a great place to work. Today, though, when I sit down with our customers, they all want to be great places to work because we're in a challenging time to hire the best and brightest and keep the best and brightest, and so everybody is focused on being a great place to work.
Actually, we partner up with Michael Bush from the Great Place to Work Institute, and we try to go into companies and show what can be done with the Workday software in the HR space. Then Michael has the data to basically give you a blueprint about what you need to do to be a great place to work. I think we are trying to be a great place to work at Workday and I think we've done very well. Our customers are now asking us how we can help them be that too. It's causing us to get more into the realm of understanding the data about what it means to be a great place to work. How do you decide where to draw the line in terms of investment? Obviously, you spend money. You spend resources on employee engagement, employee happiness, but there's a cost. That money is coming from someplace else.
I don't know if there's a cost because if the result is you get the best and brightest people to join your company and you have lower turnover, maybe it nets out. The cost of turnover is much higher than most people pay attention to. There have been studies along those lines for years. Turnover is really expensive. You have somebody in the company who has been with you for two years, now adding real value, and they leave. That really hurts because then you have to replace that individual with somebody who is starting from scratch again. Turnover is really, really painful and really expensive, both in hard dollars but also in the knowledge that walks out the door. I think, by creating the environment around the values and culture that we have, we have lower turnover. What advice do you have for companies who are listening and they want to have that higher level of employee engagement and employee happiness? Summarize everything you know on the topic. [Laughter] First and foremost, the thing is to survey your employees and for the employees to believe that those surveys are actually going to be acted upon. I think it's really important that those surveys are largely done anonymously.
You don't really get great feedback if they're not anonymous. You can't just do it once a year. You've got to do it consistently. That feature is in the Workday product, the pulse surveys. That's number one. Number two, it's creating transparency and communication channels that everybody in the company feels like they're part of a bigger mission. Transparency, in many ways, is at odds with all the rules of being a public company. I take the risk of leaks because I think the risk of not being transparent is way more costly. Maybe not around financial information on a quarterly basis. That is something that we're transparent with after the quarter is announced but, pretty much everything else, we try to be transparent and open. I think that goes to a great length for the employees to feel like they're engaged and part of building something more important.
Then I think it's having a reward system that rewards your core values. If you have a set of core values but there's no reinforcement of those core values, they're really not that valuable. We talk about our core values. As an example, we have innovation as a core value. Every quarterly meeting, we reward the Innovator of the Year award, and we have a number of those. It's not just engineering. It's across all disciplines. People see, hey, if you do something innovative, you'll get rewarded. We have the team awards every quarter. It emphasizes teams that are across the company, not just in one function. It creates a behavior. Whatever you pick as your values, make sure that you have ways to recognize the values and reinforce them. In the case of situations where somebody who might be really talented is not living up to the values, you have to take action. In the past, we had a very talented salesperson join us and we were very clear, "Don't bring any data from your previous employer." Well, this individual did. We took action immediately because if you don't, then why do you have core values? Ultimately, you have skin in the game when it comes to employee happiness, employee engagement, and adhering to your values.
Yes. Finally, my last question is back to the machine learning topic. What advice do you have for business leaders who are looking at machine learning and they want to do more with it? The first thing I would do is read that book Prediction Machines. I think it's a great read, it doesn't take a long time, and it's not necessarily about the technology. It's about the business value of better predictions and how machines and humans have a symbiotic relationship going forward. I think it's just a great read and it simplifies a lot of the mumbo-jumbo that's out there and some of the hype. I think it's good to get grounded that way. Then I would encourage folks to pick a couple of projects to try it out and see the benefits. I wouldn't necessarily recommend going all-in, in all areas at the same time. Pick a couple areas and, maybe if you're in HR and finance, go figure out where machine learning is being used in other parts of your organization.
If you're, again, an insurance company, we have a lot of insurance companies who are both HR and finance customers. Somebody in the claims world is doing machine learning. I guarantee it. Go sit down with those individuals and learn about how the technology is being used. Learn about it. Take a measured approach but go forward and do it. Absolutely. We have been speaking with Aneel Bhusri. He is the co-founder and the CEO of Workday. Aneel, thank you so much for being here with us once again. Thank you, Michael. Always a pleasure. Everybody, before you go, please subscribe on YouTube and hit the little subscribe button at the top of our website and subscribe to our newsletter. We have amazing shows coming up. Check out CXOTalk.com and we will see you again next time. Have a great day.