 Welcome back, everyone, to theCUBE's live, because we're here on the show floor, SAS Explorer. I'm John Furrier, Dave Vellante, your host of theCUBE. Getting all the action, AI analytics at scale, developer first, then rise of the data developer, a lot of action happening here. We've got two great guests, Udo Skravo, SAS, thanks, President of Advanced Analytics. I got it right. You got it right. Great name, Alice McClure, SAS Director of Product Marketing. Thanks for coming on theCUBE. So first of all, congratulations on great keynote. You had the demo sections were phenomenal. Tweeted out that twice. Alice, a lot of product messaging around AI, a lot of greatness coming on the product side. AI's been a tailwind for SAS. I mean, you guys just, it's been a gift for you guys with all the experience you've had. Take us through how you guys are feeling right now. Oh, you know, I think it's a bit of a bread and butter space for us, like you said. It's a tailwind and it's something that certainly we're well positioned to help customers. And we've been well positioned to help for a lot of years now. So this generative AI influx and the hype around it that's been happening in the market over the past six or nine months now, I think we're very excited to kind of cut through the hype of it all and help to make it real this week. We do a lot of shows, obviously. And for those that really didn't have a heritage in AI, they come up with a good story, good marketing, good tag lines. Reference architecture. Yeah, and that's fine, but so how would you describe the way you guys think about AI and maybe give us a little insight as to your AI journey? So the way I like to think about is that AI is just another, in quotes, technology which enables us to solve the business problems of our customers. You know, and this is how I think we separate ourselves from most of the other vendors where there's technology first and then the technology is in search of a problem. We like to think about ourselves like, you know, with the vast amount of customers we have, they approach us with their problems and say, can you help us with that? And now AI is another very powerful tool in our tool set to help them address their business problems. So if you're wondering about the journey, you know, I believe initially it was all about low-code, no-code interfaces, people were lacking the capability of writing code themselves. This is where we started, I believe, 10 years ago to create what we now call the Viya Enterprise ecosystem. And interestingly enough, the wins have changed again, right, so now customers are approaching us and say, can you please give us a lightweight programming environment? We are not so much interested in the low-code, no-code environments, we just want to code away and you know, I think this is why we've released Workbench. The Workbench is actually a nice addition to the Viya Enterprise, if that makes sense. You know, one of the things we've seen in the history of the big data movement past 13 years has been great ideas, but when you have put it into practice, getting the developers is key. And then two, operating it, how much does it cost to run it? And then, you know, it's hard to get that skill set, they all got hired away by all the hyperscalers and then it costs a zillion dollars to run a bunch of clusters. How do you operationalize a model once you have it? You know, how do you put it into the hands of users who are not necessarily experts in data science or in analytics or in statistics? I think the example which we saw today from Cambridge University is a prime example of how we envision this in the future. Explain real quick. So, Cambridge University approached us with a desire to match kidney donations to patients who are waiting for a kidney transplant. And in the past, they had to look at pictures and basically try to understand, is this kidney a good match for a certain patient? Now we are using image analytics, AI, to analyze those pictures and give recommendations to the practitioners. So, the practitioner, of course, is not a data scientist, but they will basically now depend their decisions on the suggestion of the AI system. And just for the folks that didn't see the keynote, this has really highlighted the key news that you announced this show, Workbench and App Factory. And you had a nice demo, not demo, but a role-playing employees, back-end engineer and a front-end engineer, not knowing each other, not kind of competitive, kind of fun to watch, but it's highlighted how easy it was from a cross-discipline standpoint to work together, kind of asynchronously and just magically, just with a line of code, integrate multimodal capabilities, images into the code. That's kind of where the direction's pointing to. It's very powerful. I think what we like to talk about is take the computer science out of the data science. And what we mean by that is, you know, as a customer, you shouldn't waste time thinking about how do different technologies fit together? How do I move from A to B? And then you have all these challenges of different APIs, blah, blah, blah. This is where I believe we have a very powerful environment where we can basically allow our users to focus on solving the business problem and we take care of the computer science for them behind the scenes. So the mental model we've been using is Uber for the enterprise. You talk about operationalizing AI, and I want to test this with you guys because you're right at the heart of it. Do you think about Uber? It's people, places, it's drivers and riders and ETAs and transactions and money and, you know, goods that you're delivering and all these data elements, they make coherent, which is what you guys do with your, I guess your semantic layer. So, but we see the future of data apps as that, you guys talk about digital twins. It's the digital representation of your enterprise. Data's coming in, you're taking action. Is that part of the product strategy? Is that technically, maybe not feasible today because you don't have to have 3,000 Uber engineers to write this stuff, but businesses ultimately want that horizontal platform so they don't have to think about the AI. It'll just be operationalized for them. Is that, in part, the strategy? Is that feasible? I love your analogy, and I think it's representative of the strategy. You know, I think a big part of this and with the workbench and with App Factory, it's all about, you know, these are software-to-service offerings. We're wanting to make this a lightweight, accessible environment. We want to create a sandbox kind of mentality for developers to be able to be innovative and to be able to do their work in an easy way and also have access to the data in a very straightforward fashion. So, I like your, I love your analogy in that it's very user-centric and very kind of persona-centric based on what's the work you need to do and how does that fit within the total scope of what's happening across the enterprise? It brings up the diversity too around the data that's out there, the databases, real-time streaming versus maybe historical data, which is now great now for with AI. You can actually look at historical stuff working together in real time. Knowledge graphs. That's a hard problem. And there's a lot of legacy databases, a lot of plumbing and a lot of systems out there. And you may have been surprised to see that we're also working on generative AI for simulation of data, right? So we want to be able to create what we call synthetic data. So you guys have been in the big data world for quite some time. So you may wonder, but don't we have enough data? Why? Explain synthetic data for the folks so we can capture that. Synthetic data, the way to think about this, you remember the deep fakes. You look at a picture and all of a sudden you realize this is not a person. This is an image which was created by a computer which is representative of attributes of us all, right? But that person does not exist. It looks very real. So we thought, well, can we apply the same concepts to tabular data? So why would we want to do this? Like I said, you know, big data, you have enough data, why are we creating even more data? Well, the point is, when it comes to relevant data for building AI models, we are actually still starving for data. I'll give you an example. Imagine you want to detect fraud, right? So now if your database is full of fraudulent customers, you are out of business, right? You know, you don't have a business. So typically your fraudulent people in your database is a representation of 1%, maybe 2%. AI models try to detect or build a model which detect those 2%. Now with synthetic data generation, we can basically now artificially increase the amount of fraudsters you have in the database because we control the data. So the relevant data can be increased which makes it easier later on for the AI model to build a model to detect actual fraudsters. Does that make sense? Yes, it makes a lot of sense. It reminds me of Tesla who sends very little data back to the very small percentage of data. I mean, maybe a deer runs in front of the car, they capture it and send that back, but there's a lack of those incidents, right? Because most of the time, the Tesla is driving just fine. And that's what a model completely not interesting, right? You know, it's like, yeah, if everything went fine, why do you need a model in the first place? It's this instance where something is happening where we don't have enough information, but that's exactly what we want to model. That makes sense. Yeah, and this is the advantage of models. This is why the model management, I thought was a great message you guys had there. Kind of underplayed, but model management now is a discipline. Oh, yeah. And a power law, as we reported on our analysis, we put out the power law of foundation models. You could be big models to small models, and they're all working together. Yeah, we call this deaf ops. I think that's just the terminology which is used. And the reason why this is growing in importance is that where's the availability of very powerful AI models and machine learning models with the availability of automation, where you can basically press a button and say, build a model for me. Well, all of a sudden you have an explosion of models. So how do you keep track on which models are still relevant, which are still giving you the answers, which you are actually wondering about, and when should you replace those models? And this is where model management is coming in big time. Yeah, and those three pillars that you talk about, synthetic data generation, digital twins and LLMs, they're part of a, am I correct that they're part of a virtuous cycle? I create a digital twin of my business, I then create more of that sparse data, and then I can use LLMs actually to feed that. Yeah, yeah, spot on. Yeah, no, not meant to be independent pillars that are completely different use cases. So coming in and out of those three is certainly a critical part of what customers are looking forward, making use of synthetic data as they need to to train models and leveraging large language models all together. Alice, how hard is it to tell the story these days? Your customer base is diverse, got a lot of customers, been working with you guys for a long time. You must have buckets of customers, people who are leaning in hard, and the early adopters are out there, then you got everyone waiting for things to shoot to drop, they want reliability, they want pragmatic solutions. How do you tell this story? Because, I mean, it's a whole other world. You got basically abstracting away the complexity of legacy and existing data systems, giving workbench and app factory the window into that environment, which is going to probably spawn a tsunami of native AI apps for your customers. How do you get that out of the way? We have such, we're in such a unique position here at SAS. So we have 45 plus years of customer interactions. And we, there's some customers that have very straightforward needs, right? Just very kind of simplistic needs when it comes to the technology that we offer and then we have extremely diverse, complex customers who are undergoing massive transformations and they have thousands of models in production at any given moment. So when we're addressing such a diversity of customers with net new technologies, and there's such a hype around AI right now, it's part of why I got into technology in the first place. I think that's why we're all here. It's an exciting thing to do, but knowing that we ultimately, we do have a capacity to help in lots of different ways. And from a solutions perspective, from an industry specific engagement, I think that's where we can help to make a lot of the hype more real. It's interesting. Dave and I were talking, I forget what event it was. We're saying the hype cycle, we were saying to each other, it's the first time we've seen a hype cycle match the value with it. And the adoption. And the adoption is more action. We have the data from our survey data. There are people are not just kicking the tires, they're actually implementing. So you have the acceleration of product market fit on the product sides for the customers. The technology's new and net new game changing. And that's happening fast. And then they have to evaluate how to deploy it. I mean, that's a customer standpoint. And how to govern. That's hard. It's hard. On their side. Nevermind, your side. It's hard. And like Alex explained, we have a wide range of customer. We like to talk about the analytical maturity. And some of them are just at their starting point. They are still dealing with data. They create some reports and it's good enough for them. But we have customers who do really exciting cutting edge AI technology, where we learn a lot from them. How they deploy and use our technology to solve amazing, amazing problems. This is where we are actually benefiting from those customers as much as they benefit from us. How about the Viya sort of progress, maturity, roadmap, where have you come from? What are you excited about from these announcements this week and give us a little glimpse of where you're headed? Yeah, I'll start. So I think as we think about again, the spectrum of customers that we have and the Viya platform, making that direct application and that direct applicability to our existing customers is core to what we do and what we're trying to convey. So for example, the enterprise guide. Enterprise guide is a product that we've had for a very long time. Lots of customers use it and love it. And that is a direct development environment within the workbench product that we're bringing to market and that's really important for us to have that direct applicability in addition to VS Code and Jupyter Notebooks and whatnot. It's an example of how we're taking the Viya platform and making it applicable to what customers are using today. Yeah, you've got a lot of cheers in the oven. It's probably the biggest cheer in the audience. Yeah, that's it. Yeah, that's it. There's a lot of people using that. Yeah, there you go. What's the big news that people should walk away with obviously the Viya platform is great. What, if you had to stack the top three most important stories for people to pay attention to from day one, what would you say is the most, three most important points? Well, in my mind it's the launch of the Workbench and Act Factory. I mean, that's number one because we are basically going after a market segment where we were strong in the past but we have not paid enough attention and now we are basically going after that segment very aggressively with a very unique technology. I think this is big, this is unique in the market and I believe this is one of the main stories which I would love for our attendees to take away on the day. Number two, our strategy when it comes to Generative AI, that we are not aiming at building our own large language model, yet another large language model who cares, right? But that we have a broader strategy, the three pillars we talked about, digital twins, synthetic data and LLMs to solve business problems. I think that's number two. And number three which may sound odd to you our customers need to understand that the SaaS you are talking to today is not the SaaS which you used 10 years ago, 10 years ago, maybe at university. We have revolutionized the way we do software, right? I don't know how many discussions I have with people like, oh, I used SaaS at university. I did too, and I was like, this is not my father's SaaS. So I think that is one of the points which we are trying to get across, like we are not the same SaaS like you did like 15 years ago when you wrote in a program editor. And that's why we're so happy to be talking with you guys today, because this is, it's really, I think of all the three, I think the third one's most important is that there's been a lot of evolution. And many folks know SaaS as the programming language and that's great, and there's so much more. I think you guys did a home run here. I think in these markets like AI where there's a major inflection point, cool and relevant works, and being cool as AI, relevant is being pragmatic about what the value proposition is. If the hype goes away, the value has to emerge. At the end of the day, that's what's going to be judged by market, I mean, the value you bring to the table. And the enablement. Here you go. It's innovation though, it comes back to innovation. We've seen so many companies in the history of this industry that stopped innovating and became irrelevant. You guys kept investing and innovating. Maybe it's because you've been a private company, you're able to do that without all the ebbs and flows of the 90 day shock clock, although, hey. Well, we know we can't rest in our laurels, that's for sure. Maybe you guys should go public. Okay, all right, it looks like we're done with this segment. All right, seriously, thanks for talking about it. And congratulations, good demos, love the demo. Demos is a new gold standard on AI at Benchmark. Who's got the data? Who's got the, who's positioned well? That's the question everyone's going to be asking. As the data revolution comes with AI, do you have what it takes? Are you prepared? Are you leaning in? Is it a tailwind or a headwind? Of course, theCUBE will dig in. I'm Sharper Dave Vellante. We'll be right back with our next guest after this short break.