Chief Data Officer
Kevin Bates is the Chief Data Officer of Fannie Mae. Kevin joined Fannie Mae in 2004 and has grown with the company through successful delivery of technology and data solutions.
How Can an Enterprise Gain Value from a Cloud-Based Data Ecosystem?
SEPTEMBER 14, 2023
In this episode of The Next Big Question podcast, Chief Data Officer Kevin Bates of Fannie Mae leads a conversation about how enterprises can gain value from a cloud-based data ecosystem. He discusses the importance of a strong data foundation and the “personas” that should inform the priorities for enterprise leaders. He also touches on generative AI and machine learning and how they can fit into this model.
Liz Ramey (00:12) Welcome to the Next Big Question, a podcast with senior business leaders sharing their vision for tomorrow, brought to you by Evanta, a Gartner Company. Each episode features a question with C-suite executives about the future of their roles, organizations and industries. Thanks for listening. I'm your host, Liz Ramey. Now let's hear what today's next big question is.
In this episode, I get the opportunity to speak with Kevin Bates, the chief data officer with Fannie Mae. I asked Kevin the question, How does the enterprise gain value from a cloud based data ecosystem? Kevin takes us through his vision and journey of a high functioning data ecosystem and the different personas across the organization that will gain value from this highly available data. He also spent some time talking about opportunities in the AI space that depend on these types of data environments. Kevin Bates, welcome to the Next Big question.
Kevin Bates (01:14) Thank you. Good to be here.
Liz Ramey (01:16) I am thrilled to be talking to you today about cloud based data ecosystems. So, well, let's just jump in as a financial institution at Fannie Mae, you’re most likely advanced in this kind of development of your data ecosystem. So, can you tell me a little bit about that journey and what stage of maturity you would say you sit right now?
Kevin Bates (01:37) Sure. Yeah. To your point, as a financial company, as a very large market maker, heavily regulated organization, we have spent many, many years focused on our foundations, in particular around our data and data management. So, over the years, that focus has really been primarily about creating connective tissue with our data. What is a loan? What is a security? Do we talk about those the same way across all of our different systems? Do we model our data the same way so that we can very quickly, when we have a new activity or a new initiative, we know how to bring the data together and to make it useful for the business to tackle that new thing. So, as I mentioned that involved those foundational efforts involved identifying the right technology and tools to help us manage the data, as well as driving that kind of adoption into our business community and users. So we like to talk about it as the three big hills. That first big hill was establishing those foundations, the right technologies, the right language for communicating about data at our enterprise. The second big hill was around driving adoption of that, moving it to the cloud, which has been something we've been focused on over the last three years or so. And then the third hill, which I think of as kind of the last mile, is bringing those foundational capabilities that are now transformed and they're agile and modular, bringing those into our analytics and modeling worlds where we need to move quickly along with our business partners to enable them. So, it's been quite a journey.
Liz Ramey (03:24) Yeah. And I love that. I love that kind of visual of three big hills, right. And kind of chunking that journey out to get to that North Star where you want to be. If organizations were to model this kind of foundational build, what sort of value would they expect to gain? And then on the other side, if they don't build these foundations, what are they going to miss out on?
Kevin Bates (03:51) Yeah, And you know, from our perspective, having been, you know, talking with our management about these initiatives and the level of investment needed. It ultimately comes down to being well-managed as an organization around your data. It’s the lifeblood of the company. It's very important for a regulated company in the financial industry. It has to be ugh. It has to be, right? So, there's no one formula for what well-managed means. But you know it when you have it as a company. And ultimately the benefits you're looking for are the ability to move quickly, to move strategically, to not have to reinvent your foundation every time you have a new effort that you need to support. But to actually see more of and and creative kind of building of the foundation where the new efforts not only leverage what you have, but maybe help you make it better. And so that sort of convergence and acceleration is what we look for, it’s something we benefit from with every new, big initiative that comes up. And so it really pays off over time.
I think on the flip side, if you're in a position that you haven't been able to make those investments or you started off on a five-year journey but had to stop at year two to focus on something new, we all know what that feels like. It ultimately incurs a lot of legacy technology debt that you have to drag along behind you. It just makes it harder. So, it's a bit the opposite of the benefits you're looking for. It becomes it's increasingly difficult to do new efforts. You have to tie a lot of things together with, you know, chewing gum and tongue depressors or whatever it is. And we know what that feels like. And that's often the call to action that you really need a take stock of where you are and see if you can address some things foundationally. So, when you feel that Frankenstein like technology and so many manual processes to bring it all together, that's usually what it's what you don't want to don't want to be.
Liz Ramey (05:54) Yeah, it seems like, you know, not to overplay the metaphor, but you're not going to be in shape for the last mile if you don't climb the first two hills.
Kevin Bates (06:05) Yeah, I think you're right. I think you're right. You're not prepared.
Liz Ramey (06:08) Yeah, exactly. So you and I have talked a little bit in the past about this idea of personas and the different personas across the enterprise, including this the C-suite and other leaders. Can you explain to me a bit about this concept and how data teams should be driving to support these personas?
Kevin Bates (06:28) Sure. It is one of my favorite, I guess, structures that I like to use to just frame my own activities and make and try to try to convince myself that I'm not forgetting something. It's from an enterprise perspective, looking across the company, you have different groups and they have different needs and you can call those personas. But, you have, an organization is usually built around those different personas, and what they need is often functionally different or used for different purposes and expressed in different languages. But a lot of times from a data perspective, it's pretty much the same thing.
We work as data professionals with the Chief Information Security office. They're going to be focused on data protection, on data loss prevention. We work with the Chief Risk Office and they're focused on how do you manage risk around data and how does the risk around data flow into risk around closing your books or risk around modeling or other things that are so reliant on the data that feeds into them? Obviously, the chief financial office is often the final arbiter of data quality and data completeness because of Sarbanes-Oxley and other things that many of us in the data space and the financial industry deal with.
So, I think ultimately these personas, oh and by the way, I shouldn't I shouldn't miss the business is a very important persona that use the data to actually drive our business forward. And they need what they need, they needed at the speed that they need it for new features, new capabilities. And I think if you take a well-rounded view of all of those and say these are the people that inform your priorities as an enterprise leader, that that's really going to serve you well and you'll often find a common language among them where you can harmonize what you're doing to support all of them at the same time. But it starts with being aware of what they need and making sure that it maps to what you're doing.
Liz Ramey (08:34) I guess like painting the picture here, What would you say are the qualities of a high functioning data ecosystem? And then and to further that question, how does an organization know they're ready to really trust the data and decisions that are being made when you're in that type of environment?
Kevin Bates (08:53) Yeah, and that's sort of the question, you know, we all have every day. Is our data good? Can we trust it? I'll hit a little bit of the same territory I covered. I mean, I ultimately think we all speak for ourselves. We're in different organizations with different priorities, but a heavily regulated financial organization, it's really going to come down to controls. You have to have the right controls in place so that you're confident that you're measuring things at the right points in time, such that if the control fails, you know, it would be problematic. If the control succeeds, it will catch the things that would be a problem before they happen.
And so I think the ability to respond quickly to new requirements while having controls in place that help you safeguard your business, that's really about safeguarding your business process. That's what you need. So how do you trust that? I think it is ultimately measuring what you manage, managing what you measure, having early warning indicators in place so that you can see. I've had, for example, I see a flurry of data quality issues in this area. There may be a technology issue and that could over time become an issue for our business downstream. And so those controls are the things that alert you and those and you identify some of your controls as being an early warning indicator and other controls to be essentially where, how you're operating.
And if you have metrics that you can show others that you trust yourself, you're managing to them, you know what they mean, and you can share them with your chief risk officer or with your chief financial officer. Then I think you'll build confidence across the organization that to the extent that that you understand that you are managing data effectively and transparently. I think when you have that, that really that will help carry the day for you to trust your data, to trust that you know how to make changes to your data and measure that everything's okay. So that is a data management practice 101, I guess I would say. And if there's something where you have to be transparent and grind the ax against those controls over time.
Liz Ramey (11:13) So I can't have a podcast with a chief data officer and not talk about AI or large language models. So I have to ask, you know, let's talk about this. Let's talk about these kind of large language models, the different forms of AI, generative AI and others that are out there, and they're pretty hyped right now. And so if an enterprise wants to take advantage of these, what do they need to have prepared and what should they be prepared for? Assuming that they're building these kind of cloud-based data ecosystems in order to build these other capabilities?
Kevin Bates (11:50) Okay. Yeah, absolutely. A lot of energy in this space for sure. I can give my perspective on it because I am very much in the middle of what our company is doing when we think about generative AI in general. One thing that's worth noting is AI and ML, machine learning, all of those are techniques that have been around for many, many years and many of us are using them. What's new now is this concept of generative AI and these large language models, which in their most powerful state exist in the Internet. They're out there in the world and they're learning from everything that's happening in the world, Right? And that's not how many companies can operate.
They have to have real control over what data they use to inform their business. And so if you, I think all of us are in a position that we want to take advantage of these capabilities. They're exciting and they can be game changers. We see a couple of different variations on that. Obviously, you have coding, code generation, the ability to make your developers much more efficient, enable them to translate, you know, coding languages from one stack to another stack, and that's a very powerful capability. That's a big part of what folks are thinking about nowadays in terms of generative AI.
And then there's sort of separate, you know, separate from that business transformative ideas where you could do things like take 200 gigabytes of unstructured data and summarize it in a Word document completely kind of automagically. So there's a lot of opportunity there. And with those opportunities and I mentioned a little bit about how large language models work, there's also a lot of risk. So every organization is different. From my perspective, there's the balance of risk, and opportunity requires us to be very controlled and to have essentially to pilot out these new tools to see how mature they are. How would you really integrate it into your company? Do the existing controls you have protect you from the downside? For example, we don't want to have our code put into the Internet. That would be a data loss event for us. It would be a really big deal. So we have to understand how do we control it and use it all for what they're good at while keeping a human really at the center of our business decisions. So that's a bit of work. It requires us to move cautiously, but we also want to move forward and take advantage of the capabilities.
Liz Ramey (14:35) Yeah, it's such a balance right now, right? Because we talked about this kind of getting your environment and then your organization at the point where they can trust the data and you can't really trust the output of these models yet.
Kevin Bates (14:52) Yeah, it's a very a lot of a lot of risks to manage. Trusting data is hard enough in the first place. To your point, it's well known that one of the biggest risks with these, like a ChatGPT or these other kind of models is that they'll give you very confidently wrong answers if you don't have a human and a control at the heart of the process, you can't trust it for business. My $0.02, including just generating a bunch of code. If you don't have a developer who's accountable for every line of code, that could be a risk for you and your organization. So it requires a fresh look at our controls, potentially some new ones, and it requires us to really look at like, what are we trying to achieve? What is our goal? You know, Fannie Mae is a regulated financial services company. We have different needs than a company that may be focused on call center operations and things like that. So we have to understand what we're doing, why we're doing it and how are we controlling it.
Liz Ramey (15:55) All right. So we've come to the last two questions, and one is from a former guest, and I'm so curious to get your thoughts on it. And then I now ask you your final question. So this is from Deborah Wheeler the CISO of Delta Airlines, and she asks, What do you think the relevance of the metaverse is going to be on our society going forward?
Kevin Bates (16:21) Well, thank you. So that's a great question, and I thought about that a little bit myself. I think we've all probably been evolving rapidly on this topic. From my perspective now, Metaverse, virtual reality, these are real. They're fully up and running in some industries and are at the heart of some industries like, you know, video game gaming industry and there are others. Right. And and so in some areas this has actually been used for a long time and it's now just maturing into relevance in other industries, I would put it that way. I think as with many technologies, what's table stakes in one industry is like this brand new concept in another. So since we've already transitioned into a metaverse-like state, we've increasingly become reliant on, you know, collaboration tools that allow us to interact from all over the globe. Global companies do this. The pandemic, I think, really drove home that we just get comfortable with remote-only work. And in a sense that puts you out in the virtual world already. Bringing in a VR headset to the equation becomes more of a question of do you have a good business need that a VR headset or a virtual physical reality, a virtual reality to go along with your physical, physical reality really match up.
So I would say that we are essentially there. It's knocking at the door. A company that is a heavily regulated financial company might have less tangible, immediate need, but like a little bit like what we talked about with generative AI, very soon you realize that this is an efficiency tool or it is a tool that allows you to be more productive and you need to adopt it. So you can't just ignore these things. You have to bring them in. So I think that I think ultimately the metaverse is here to stay. We will see business use cases pop up where, it's not science fiction anymore, it's something that we can literally leverage seamlessly, especially now that we're all used to working in a remote sense. It could give us new ways to collaborate across the Internet and across the world. So I see it's something that we have to continue to keep our eyes on.
Liz Ramey (18:48) So finally then, Kevin, I'll have a guest on after this episode. I'm not going to tell you what C-suite it will be because I want this to be a broad question, but I do have to ask for you what is your next big question that you think executives should be pondering right now?
Kevin Bates (19:09) Okay. we've been spending so much time on generative AI and these tools, and those aren't just data things, those impact every aspect of your organization. And so what I thought was instead of thinking, you know, talking more about just generative AI and what it is and what you could do with it, think about it. Take a step back and think about, as a leader in your organization, how are you preparing your organization for these kind of game changing events? The next big thing, if you will? So we're currently dealing with Gen AI. It means that we have to have a collaboration across the risk community, the finance community, the business. We have to have communication strategies and other things. So then when you look down the road and say, okay, this probably won't be the last one, maybe the metaverse will be the next one to take off in a really big way. How are you preparing your organization for that change? How are you being a change leader? And where do you see the capabilities being framed appropriately or inappropriately? How can you be the one to help the organization understand the opportunities and the risks and lead them to kind of the right place based on what they do? So I think that's something for all leaders ultimately to think about their role in that. And so I would put that forward to the next person on your podcast.
Liz Ramey (20:39) It's a great, that's a great question. I'm going to have to think about that myself. But Kevin, thank you so much for joining the next big question. I really appreciate.
Kevin Bates (20:46) Thank you for having me. It was a pleasure.
Liz Ramey (20:49) Thank you again for listening to The Next Big Question. If you enjoyed this episode, please subscribe to the show on Apple Podcasts, Spotify, Stitcher, or wherever you listen. Rate and review the show so that we can continue to grow and improve. You can also visit Evanta.com to explore more content and learn about how your peers are tackling questions and challenges every day. Connect, learn and grow with Evanta, a Gartner Company.
by C-Level, for C-Level
Find your local community and explore the benefits of becoming a member.