Jay Upchurch, CIO at SAS Institute, joins theCUBE’s Scott Hebner and Paul Gillin at SAS Innovate to explore the strategic value of composite AI in the enterprise. Their discussion highlights how organizations can move beyond generative AI to embrace purpose-built models that drive smarter decision-making and competitive advantage.
Upchurch offers a grounded take on the role of composite AI, sharing how SAS combines multiple techniques to solve real business problems. He emphasizes the importance of aligning AI tools with enterprise needs, and explains how thoughtful integration is key to meaningful adoption at scale.
The conversation also touches on change management and the evolving role of IT leaders in orchestrating AI strategies. Upchurch and the analysts examine how to balance cutting-edge innovation with readiness, offering practical insights for organizations navigating the next phase of AI maturity.
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
SAS Innovate 2025. If you don’t think you received an email check your
spam folder.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open this link to automatically sign into the site.
Register For SAS Innovate 2025
Please fill out the information below. You will recieve an email with a verification link confirming your registration. Click the link to automatically sign into the site.
You’re almost there!
We just sent you a verification email. Please click the verification button in the email. Once your email address is verified, you will have full access to all event content for SAS Innovate 2025.
I want my badge and interests to be visible to all attendees.
Checking this box will display your presense on the attendees list, view your profile and allow other attendees to contact you via 1-1 chat. Read the Privacy Policy. At any time, you can choose to disable this preference.
Select your Interests!
add
Upload your photo
Uploading..
OR
Connect via Twitter
Connect via Linkedin
EDIT PASSWORD
Share
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
SAS Innovate 2025. If you don’t think you received an email check your
spam folder.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open this link to automatically sign into the site.
Sign in to gain access to SAS Innovate 2025
Please sign in with LinkedIn to continue to SAS Innovate 2025. Signing in with LinkedIn ensures a professional environment.
Are you sure you want to remove access rights for this user?
Details
Manage Access
email address
Community Invitation
Jay Upchurch, SAS
Jay Upchurch, CIO at SAS Institute, joins theCUBE’s Scott Hebner and Paul Gillin at SAS Innovate to explore the strategic value of composite AI in the enterprise. Their discussion highlights how organizations can move beyond generative AI to embrace purpose-built models that drive smarter decision-making and competitive advantage.
Upchurch offers a grounded take on the role of composite AI, sharing how SAS combines multiple techniques to solve real business problems. He emphasizes the importance of aligning AI tools with enterprise needs, and explains how thoughtful integration is key to meaningful adoption at scale.
The conversation also touches on change management and the evolving role of IT leaders in orchestrating AI strategies. Upchurch and the analysts examine how to balance cutting-edge innovation with readiness, offering practical insights for organizations navigating the next phase of AI maturity.
Jay Upchurch, CIO at SAS Institute, joins theCUBE’s Scott Hebner and Paul Gillin at SAS Innovate to explore the strategic value of composite AI in the enterprise. Their discussion highlights how organizations can move beyond generative AI to embrace purpose-built models that drive smarter decision-making and competitive advantage.
Upchurch offers a grounded take on the role of composite AI, sharing how SAS combines multiple techniques to solve real business problems. He emphasizes the importance of aligning AI tools with enterprise needs, and explains ...Read more
exploreKeep Exploring
What are some considerations and challenges in adopting artificial intelligence for business enterprises?add
What is the risk of getting distracted by new AI technology and neglecting basic IT functions and business goals?add
What are some examples of how SAS is utilizing AI technology in their operations, specifically in the context of their campus in Cary, North Carolina?add
What are the three things that make SaaS special?add
>> Welcome back to theCUBE's continuing coverage of the SAS Innovate 2025 conference. I am Scott Hebner. I'm here with my colleague, Paul Gillin. And Paul, we got to talk to a lot of people today.
Paul Gillin
>> We sure have and it's been fascinating.
Scott Hebner
>> A whole bunch of different perspectives on what's all going on here and it seems to all come together well. What we're going to do today is bring in a CIO perspective and I'm happy to introduce Jay Upchurch here. Welcome to the show.
Jay Upchurch
>> Thank you. Thanks for having me.
Scott Hebner
>> Yeah, and I think we'd like to hear a little bit about the more holistic view of what's all going on here, this notion of composite AI, and maybe you can start by just defining what that is.
Jay Upchurch
>> Well, let's take a step back for a second. The struggle, I think, for all of us is very real for adoption of AI, how to do it, the right pace, the right sequence. And then, I'd also say the opportunity is very real. So, when we think about, again, the competitive advantages that a company can gain by deploying artificial intelligence in a meaningful way to meet your business enterprise needs is a great differentiator for everybody. Actually, composite AI, the way that we think about it, that there's not really just one AI technique that meets the needs of everything. Earlier in the keynote today, you heard Bryan Harris talk about sprinkling the pixie dust of generative AI to solve all your world's problems. And when we run an enterprise, like for instance, my job at SAS, I have to think about fit-for-purpose. What's the right AI technique? Remember, gen AI is just one of many that you have in your toolbox. You may use computer vision as an example. You may want a different type of artificial intelligence technique to meet the business need, again, to hopefully achieve your AI ambitions.
Paul Gillin
>> So, is composite AI an architecture or is it more of a state of mind?
Jay Upchurch
>> State of mind. It's philosophy at the end of the day. I think what we saw in the industry, call it the last 18 months, we were all enamored with what the capabilities were for gen AI. And so, we started trying to take generative AI and use it to solve every problem that came in front of you. And so, we spent a lot of money, unfortunately, as an industry chasing problems that were actually probably better served with other types of artificial techniques. So, again, back to this idea of intelligent decisioning, generative AI is going to give you an answer. It may not necessarily be the right answer. It could be 80% directionally right. In enterprise, where we need an intelligent decision made, that's not the right answer. You don't have the choice to be right or wrong there. Think for example, SAS, we have such a wonderful rich history in health and life sciences, drug development. Patients' lives are at stake with the answer that our data and our math gives you back, you don't have the choice of being partially correct there. So, those types of things, you got to figure out what's the right technique and the right moment for the business problem that you're trying to solve. So, what I think is happening, at least in the CIO communities that I'm running in, there's a lot of people taking a step back and figuring out what's the right technique, again, fit-for-purpose to meet the business need?
Scott Hebner
>> Yeah, I think that makes a ton of sense because generative AI is not necessarily designed for making decisions. It makes predictions, it's statistical probabilities, and what I'm hearing out there is a prediction is not a judgment. You need to make judgments to be able to make decisions. And so, you got to augment the generative AI with extended models and capabilities, and I think that's exactly what you're getting at with composite, right?
Jay Upchurch
>> Absolutely. We think about it the exact same way. The other thing we think about too is again, you fit-for-purpose for the needs. So, if it's an experience of an employee, a human with an assistant, generative is great. So, that's employee productivity. We've seen that now for, again, the last year. Microsoft Copilot leading it. They own the workspace that everyone is in, obviously, because they own the desktop. Next step would be, okay, now I've got humans with an assistant. And in that case it's a human leading the assistant and what they're looking for. So, think about it as I'm engaged with an agent, I'm asking it a long-range problem, it's coming back with an answer, but I'm involved in the outcome of what that looks like and I can tune the answer as I want. And then, of course the third phase, if you're walking through that maturity stage is going to be okay, now we're into agents. And that's like what we talked about earlier today around our intelligent decision capability. That agent that has been tuned and modeled, but it is doing the work for you and humans are more overseeing it, as opposed to leading it if you will, or engaging with it, just as you would with an assistant.
Paul Gillin
>> Is there a risk of whiplash? By that, I mean here today, SAS talked about four or five different kinds of AI, and a lot of organizations are just trying to digest generative. Is there the risk of taking your eye off the ball, the basic blocking and tackling of what IT should do, the business intelligence, the basic data analytics, chasing after this shiny new object?
Jay Upchurch
>> There is, and I think that's always been the plague of a CIO. We get enamored with technology. We get enamored with shiny toys. Suddenly, you start to go over there and you're making a lot of investment and then you realize say, "Wait a minute, I had a mission that I was supposed to achieve for the business." So, that's been a struggle, especially with the cost basis a lot of times of some of these new AI products. So, there is the risk, I'll guarantee that. I also think that we all are starting from a different point and that is the biggest, I think, challenge. One would say, "Okay, hey, the technology's moving at such a rapid pace, it's difficult to keep up." The other part of it is that we all are different places. What does your data state look like? Is it well-governed? Is it secure? Also, where's the direction coming from? Is it employee-driven because they want tools to do their jobs better, be more productive because they hear about it from their neighbor? Is it top-down from a CEO saying, "We're not going to do anything unless it's AI-driven"? So, I see that in our customers that we serve in our SaaS managed cloud services business, they come to us in different forms and readiness. And so, if you think about the person who's sitting on their couch and they're like, "I want to go run a marathon," you're probably not ready for that yet, might want to do some stretches and some calisthenics first. We have customers who come to us, they're like, "Yeah, we want to jump right to a gent." Well, hold on. Is your data ready for that? Are you ready for that?
And then, the last one I think back on your point around CIOs and some of the things we struggle with is the best technology with the best intentions can land poorly if you're not ready for that enterprise change management. So, what's the problems? Humans. Are we really ready? Are we going to embrace it? Are we going to run with it? The great thing about what we're seeing right now in AI in our space is that it is becoming more and more pervasive and the more people use it, the more their mind starts changing. That's not me showing up as a technologist and saying, "Hey, I got this great thing and I want you to start using it." Now, they're coming and thinking about, "My curiosity has led me to new places. Now, I need you to help me, again, deliver it to the enterprise."
Scott Hebner
>> Yeah, it's like back in the internet days, the browser gave you access to the online world. Everyone got familiar with it and businesses started very simple web presence, right? With generative AI, now everyone's getting it to experience the power of AI is a gateway into the world of AI and people are doing things with generative AI to automate tasks and help people get more information. But I think the real value is going to start to become when you start to architect around that, and I think that's where you get into all these AI technologies, which gen AI, as you said, it's just one of them and that's where the power's going to come from. But everyday people, everyday workers, they know the power of AI because gen AI.
Jay Upchurch
>> Because of gen AI. And it was targeted at us almost as consumers, curious consumers. When ChatGPT first came out, it raised the collective awareness of society and then people started trying to figure out, "Well, how do I apply that to enterprise?" And I think that's part of the reason why in the business setting, there was a lot of money spent upfront in research and chasing problems because you had cool technology, back to the shiny object problem. I think what we're seeing now is people are applying the technology in the right ways and to solve the business problem they're looking for. I think for us, again, are you a fast-follower or are you out there leading on the edge and how much is your budget willing to cover? I think that also comes back around to is it the top-down driven initiative or is it coming from the bottom up? And in our case, what we're trying to do, we're trying to take a pragmatic approach. We think about it, again, like I said, employee productivity, number one. That's an easy one. We want to make sure everybody's aware of it. I have the gift of working inside of an amazing company where our awareness and understanding of AI is probably a little higher than the average company. So, I have very demanding customers. So, putting tools in hands to raise their awareness and understanding is one thing. The next thing is, all right, how do I, again, make it where humans are interacting with agents in a way that is a long-running process that might start to change business processes? That opens up a whole other category of work. And then, the last would be, okay, now we're taking those and we're making those agents and those agents are starting to do the work that we had humans doing before. Back to this idea of, okay, what are we doing with intelligent decisioning as an example? That last step in maturity in the way we're thinking about it is really a game-changer in how you run your business, and frankly, the cost that it takes to run your business.
Paul Gillin
>> When we were talking before we started the cameras, you mentioned change management is a major goal. The barrier really is Humans.
Jay Upchurch
>> Yes.
Paul Gillin
>> Now, you've gone through a lot of this in SAS. You obviously have a technologically-sophisticated workforce, but what advice do you give CIOs for preparing the organization for the introduction of these tools?
Jay Upchurch
>> It's interesting. If you go back, we're all a pattern-matching machine, at the end of the day, the way we think. We always talked about data literacy, and then we went into, "Okay. Well, now we want to get into analytics and AI literacy." I think that's really where we're going. It's continuing to raise the collective awareness of what it is. It's not just ChatGPT where I can use it almost like a new fancier search. It is truly thinking about how I can interact and ask it questions and interact with agents and let it do the mundane work for me. And then, I love this idea that agents can do, people are concerned about job replacement, I totally get it, but they can do these jobs and free us up of thinking about other things. There was an example of a customer that we were talking to, again, back to this idea of change management, and they want to basically have agents that act like employees. And so, as a manager now, thinking about I manage a workforce of 100 people. Well, actually what if 20 of them were digital agents? And I'm tuning and training the agent to be better, but I'm still professionally managing career management, call it the other 80 people that are in my shop. Those types of changes are very real and interesting. So, then the question is, well, how do the employees handle, "Oh, okay, I see this job is now driven by an agent, but my other job now is something different"? That's where I think the industry at large is working their way through that right now. I will say this, I mean budgets are under intense pressures. We know what's going on geographically right now. I think agents and the way that artificial intelligence can come in and disrupt helps answer a lot of the challenges that we are all facing. The question is can you get there? Can you get there safely? Can you get there with trustworthy innovation as well?
Scott Hebner
>> Yeah, I mean, in my conversations, there's two sides of the coins here. One is that it's about labor savings, right? The other is about giving superpowers to your workers, so they can do more. And I have to say the superpower thought comes across twice as much as simply replacing labor. So, I'm not sure I buy into this AI is going to take everyone's jobs. Just curious to your thought on that. It seems to me it's going to be a major productivity boost.
Jay Upchurch
>> Major productivity boost. Everything that we've done in our own shop, looking at employee productivity gains through the use of things like Copilot and integrations in... It absolutely gives back, again, the gift of time to our employees. The other thing is it's beyond just a help me in my unified communications experience, wanting to take action for those employees in applications that they would normally have to stop context switch, log in, remember how to do it. Any of those kind of actions that I can take that off your plate and you can just ask your digital assistant to request that time off, enter your time entry. Anything that's kind of, again, required but mundane, let the agent do the work. That's an immediate gift of time for you. On the other hand, jobs, like total jobs where we'll take responding to RFPs as an example. You have a format that comes from a customer. You have a library of information that's unique to your knowledge domain. How do I mine that knowledge domain, give me back a generative answer that's perhaps better than what was previously done? And the human is in the loop then to review, "I've just made that employee more productive and frankly the quality of their work's higher." That is a beautiful thing in a business. And then, the last one obviously would be the full agentic side of things.
Paul Gillin
>> What are some things you're doing with AI within SAS? What are some of the cool projects you have going on?
Jay Upchurch
>> Yeah, you can imagine we have a few across our enterprise. I mentioned employee productivity. I won't go back to that one. On our campus where we run, our shop as an example in Cary, North Carolina, we have computer vision in action. We have a solar farm that feeds power to some of our buildings. Solar farms sometimes aren't as effective as you would want, whether or not it's chasing the sun or the effectiveness of a panel. So, we fly a drone, we look down at it from a computer vision standpoint, we know exactly when we need to go in and take action to maintain it. We use flywheels, as an example, for backups. We wanted to get away from battery a long time ago from an environmental standpoint. Those flywheels make noise and we can detect based on noise if they're starting to need maintenance. So, instead of paying a recurring maintenance fee, we can go in and through harmonics and figure out, "Now's the time. Let's call a technician out and come out and do that maintenance for us." So, again, back to this idea of composite AI, we look at the business problem, the use case, and come up with what's the right AI technique to meet the need? And then, we figure out, okay, how do we go deliver that in a cost-effective way, again, that meets the business objective?
Scott Hebner
>> So, let's fast forward to say it's SAS Innovate 2028, three years from now. What do you think we're going to be talking about?
Jay Upchurch
>> I don't know. I felt like we did a pretty good job this morning, trying to stay ahead of things between the quantum and obviously the intelligent decisioning and everything else that we're doing. The digital-twin capability and manufacturing was fantastic. I love the pace that we are moving at, both as SAS, but also as the industries that we serve. Predicting what's going to come next year is really tough.
Scott Hebner
>> Hard isn't it?
Jay Upchurch
>> I would tell you there are three things that make SaaS special. I think number one is we love setting the pace of innovation for this industry. We've been doing it for five decades. Number two is the customer base. Look around, you said you've talked to a lot of people today. They inspire us by telling us how they're using that innovation, and that gives us the fuel to want to go do it and come back next year with the next thing. And then, the third thing about SaaS that's so interesting is our culture. Our culture attracts the top talent. Our customers, again, give us inspiration, which feeds that talent and then that drives back into innovation. That's the flywheel of success for SAS and has been for the last five decades to me, that ultimately will help us come up with what's the next thing for 2026 for us?
Scott Hebner
>> There's no doubt the innovation cycles are moving at warp speed, right? If you think about it, it's only been, what, three years, two and a half years since gen AI came out in this space. A year ago. We weren't even talking about agentic AI. So, it is hard to look forward, but I think you've got the right idea here because a lot of technologies that are going to come together and we'll keep tracking it. All right. we need to wrap up. Any last questions you want to-
Paul Gillin
>> Nobody's mentioned robotics all day long today. Is that because it's not in your flywheel, it's not in your wheelhouse? Or is it because that technology is not moving at the pace that other AI is?
Jay Upchurch
>> It's a good question. It's probably not quite keeping up yet. I think there was a lot of early gains on it, and then I think it's slowed down a little bit. People started chasing other places first. I have a feeling it'll come back. Even in the example earlier today with Georgia-Pacific getting back around to these autonomous ground vehicles and what are they doing? You can imagine those getting much more intelligent. Again, merging in different AI techniques to make those machines a little more sophisticated. I can see it coming back around. Maybe that's a-
Paul Gillin
>> Maybe three years from now.
Jay Upchurch
>> Maybe that's coming back around 2026, that's fair.
Scott Hebner
>> That's right. Yeah. Well, we'll be back and we'll be talking about that. All right. Jay, thank you very much. This has been fascinating. Paul, once again, thanks for being here and talking with our guests. And for all you, thank you for joining us. We are the leader in tech news and analysis. Thank you for joining us. We'll see you soon.
Scott Hebner
Paul Gillin