 Welcome back everyone to theCUBE's live coverage here in Las Vegas. We're on the floor, SAS Explorer event, all things AI, data. The data revolution's here. I'm John Furrier, your host with Dave Vellante. Digging into two great guests, Gavin Dase, SAS executive vice president, Michael Galaris, program manager at Microsoft Cloud Azure with Microsoft. Thanks for coming on theCUBE. Thanks for having us. Guys, the topic here obviously is optimizing data and getting access to data you haven't had before. AI is a gift that the industry's just got that got the foundation models. It brings a lot of new things to the table that accelerates a lot of other things. Like it's like a big gravity piece of the industry. What's going on with the Microsoft relationship with SAS? You guys have been on this journey together, both individual journeys, but also together. So take a minute, Gavin, to explain the relationship you guys have with Azure and how it ties in with the AI that went great. So deep partnership with Microsoft are going on almost three years now. And it co-engineering relationship between our businesses, co-sell relationships, and really bringing transformational solutions to our customers, right? That was, our customers want to run and consume analytics, AI and data management in the cloud. And we've done, I think, pioneering work in a lot of ways to bring that forward to them. So you want to talk a little bit about kind of some of the innovations on that side? Yeah, absolutely. I think working very closely with SAS, we found that there was so many synergies where Microsoft and SAS are both going on these transformative journeys together, as Gavin alluded to, where we're really on this mission to understand the most critical industry use cases that can be unlocked and sort of pragmatically explored by industry AI, by data, and by the transformative power of generative AI. So being having, as Gavin mentioned, the two engineering teams working side by side to actually unpack those use cases, come out with prioritized scenarios and plans to illuminate the path to production, and then ultimately show our customers that here's the art of the possible, here's how you can actually accelerate, and then bring together the best of both SAS and Microsoft AI offerings together to help them own that way. So you have customers kind of rethinking their data management strategies, and now then AI just explodes. Not that it's not been here for a long time, but people become much more aware of it. How are they thinking about data management, and how are they thinking about data management plus AI, and maybe we can get into some of those use cases. I mean, for me, I came from the data management world, where I came from a company called Data Flux at SASplot in 2000, so the fact that data management's fashionable again, it's about time. And model management, don't forget model management too. Data and model management. And I think a lot of organizations were skipping to the end, right? They were focusing on the outcome, and AI and how powerful it is, but they lost their good data management practices. What data are you using it? Why are you using it? Where does it come from? What does it represent? What structure? What format, right? Like we're getting back to those things that ultimately you're going to make AI and generative AI more effective and impactful. So for me, the fact that our customers and we're having these conversations is largely beneficial. Yeah, just to pick up on what Gavin mentioned, we see a number one pattern is this idea of bringing your own data, or chatting with my own data. You know, I have a variety of different data source across the enterprise, that I want to harness this power of this generative AI to summarize and to be able to do semantic searches and the like, but if I don't have a good handle on that and I don't have sort of an understanding of where it is, or how to make citations and cross reference it to actually expedite the change management of the AI to understand that it is grounded in enterprise truth and enterprise data, then it's kind of going to be all for naught. You know, I love the data management because data management at the end of the day comes back down to quality. And also, what's what, context, right? So I love how you guys in the keynote talk about AI lifecycle, repeatability. Those kind of get me going, okay, repeatability flywheel, momentum, scale. So now comes the question of good data, bad data, clean data, all that good stuff. In a way, we're seeing AI as a generative thing, so I get this image in my head of seeds growing content, like growing data. So if you got clean data, the results are probably going to be much stronger. So is data the seeds of innovation? I think so, I think it's very well said, and I think also too, the idea that you can construct pipelines, you can construct process, you can actually bring the entire enterprise along, whether you're a line of business person, a citizen developer, a data scientist, it's all about having a new hyperscale set of tools to actually illuminate how are those seeds growing? Are they growing fast enough? Are they getting enough water? And what we also need to bring together is ecosystem partners to help uncover that and grow it even further. Gavin, if the seeds is data, then the existing legacy is the soil, right? I mean, you can get, and what I saw with Workbench and App Factory is okay, I can basically overlay infrastructure on that, extract away that complexity and get at that and grow and generate more action. Yeah, I mean, I think that's a big part of it, is lowering the barrier to entry to people to use and consume analytics and AI, right? And for me, there's also another part of AI that isn't being widely, maybe you talked about enough, but it's the use of AI in our own applications and our own technology to make that better, right? When we think about data management, eliminating some of those mundane and routine tasks so people can go focus on generating business value, that's, I think, a hidden gem from the AI world that we're starting to see emerge. I mean, that's exactly what led us to think about the co-pilot, this idea of unleashing productivity for the pilot, the human actor that's going to benefit ultimately from this newfound productivity and this newfound gains, but then whether they're offering a PowerPoint presentation or a set of new analytics and data pipelines, how do we actually bring these capabilities to accelerate and enhance what they're doing through all of our various surfaces? Well, I think Satya, I think it was at Ignite, said it really well, you're going to get a really good draft, you're going to get ideas, but you had set off camera, you can't forget about the pilot, right? What did you mean by that? Well, I mean, again, it's not autopilot, it's not replacing, you know, there's some very natural real-world concerns about displacement of the workforce with this very powerful computational capability, but what we really want to do is first and foremost, as the technology matures, make sure that there's still human expertise, humans in the loop that can look over it, make sure the answers, make sure the output, the content generated is accurate, is thorough, is ethical, is not harmful to copyrights, is not harmful to other people or discriminatory nature, and then that actually accelerates and really makes that person much more effective as that system being the co-pilot and then whether the pilot's a banker, an analyst, a developer, a lawyer, they're actually getting that benefit from that system and being able to do more further, faster, and more accurate, and cheaper. Enterprise AI is hot, obviously the chat GPT as Dave calls it, the shot heard around the world, woke everyone up, magic, it's awesome, see it, you can see it, you know it's great. Enterprises are different, and there's a conversation going on in Silicon Valley where it's a shiny new toy, and Seattle, Boston, and everywhere else, where the enterprise is more pragmatic. They're not going to just jump at implementing something that's going to have any downsides. So we see people say they're going to bolt it on first and then try to figure out where to build native AI. So I guess the question is, do you agree with that? And what are customers doing? Are they, what's their pragmatic workflow? Are they going to bolt it on to something existing, co-pilot be great, get the data, get harvest and innovation, and then if they go to the next level, what's that native application look like? What's your guys' view on that? Yeah, on the customer side, you know, we saw customers, especially in a couple of the, you know, more forward-leaning industries really rush in and we're going to start using this right now, and I think they kicked the tires with it, working with partners like Microsoft and others, and now there's the use cases being driven internally, right, 10, 15 use cases of making the business better, right, and I think that's where they're going to start to bolt this onto existing applications internally, right? I think there's education that's still needed in my opinion and a lot of the customers that we talk to, especially in the boardrooms on what this is and how it works and where they should be using it, and that's why you saw, you know, SAS, we just published AI guidelines and many others are, both for internal use and for external use, but I think it's, there's going to be, continue the flood of this throughout the, through the back half of this year, and then I think we're going to really start to see some big productivity gains in these companies in the beginning of the year. We're seeing these four fundamental patterns really starting to emerge out of kind of the initial exploratory or educational phase of generative AI, first being code generation. So I've got a variety of either proprietary in-house systems that I've built over the years or I'm working with common language frameworks like Java, Python and so forth, and I need to actually generate my code to make my developers more productive, but I also need to bring junior developers along and give them a new set of learning and education tools. Content generation, think of it as code generation, but for the marketers or for the communication staffs that need to get corporate messaging out in a much more productive and efficient way across ever more present channels, whether there's social media, publications, and the like. Semantic search, where this is that idea of, as Gavin said, you invest in the data management, now I get ever more richer search capabilities to understand what is my enterprise been accumulating all these vast years. Take the clutter, take this sort of hoarding out of the enterprise and unleash it and unlock business value. And then finally, the classic sort of conversational summarization, whether it starts with sort of a natural language question and answer, but then quickly gets into sort of this value add of being able to summarize and then cross-reference those sources of data to again really unlock productivity. So just as an observer of the industry, I mean, SaaS has been doing neural nets forever. Microsoft with open AI just totally cut the line and became number one overnight. I mean, I was sharing some of the ETR data and demonstrates what you guys have done and then of course there was this call to slow down. And you and I on our cube pod were like, yeah, that's because others are so far ahead why are you to slow down? Are people slowing down? Do you see any evidence that people are slowing down? It seems like just the opposite. Yeah, no evidence of slowdowns at all, right? I agree with your point of pretty much just the opposite. I mean, there's, our customers that I talk to on a daily basis have teams devoted to this now, right? There's active research and I think they're turning to industry partners like us to go help drive this and solve it. So no, no slowdown. And I don't think we should slow down by the way. I think that's a narrative that we don't want. I mean, part of the reason we're working so closely together is our customers are demanding it. They're demanding the ecosystem band together to accelerate and to drive standards, to drive innovation, to drive connectivity because this has to be, yeah, best practice. It has to be a team sport. I mean, there's no way, you know, we're a platform company. We have nowhere near the industry expertise and the credentials that SAS has developed over the years and we do have some things around scale and the cloud that can be better together. But that's really what customers expect. We've been staying on the queue for eight years now. You've got the horizontal scalability with vertical specialism, with the unique data and workflows. Finally now, not just forced and bolted together with some bad connective tissue, but like seamless integration. It's what the customers demanded of us, right? And for us, it was the scale of a partner like Microsoft to bring those offerings forward is what he said better together, right? That's been kind of the mantra of this partnership since the start. From Microsoft, I mean, the cloud moment from Microsoft was so transformative. Do you feel like this AI era is as big? Absolutely. I think it definitely has the potential to, because it really builds on top of that cloud foundation. You needed the hyperscale of the cloud. You needed the computational power of the cloud to really take these deep neural network capabilities, these advanced machine learning workloads and then take them to new heights and then open up the ability to process ever larger language sets. You don't get to 175 billion hyperparameters in tuning without sort of that initial cloud investment. I love how foundation models have opened up a new innovation way, because that's what was missing. We got cloud scale hitting at all cylinders, in comes the financial models, which lets you reuse stuff. You don't have to rebuild. You can do inference and training, all that good stuff you guys talk about. So everyone's enthusiastic and confident with another discussion. But let me ask you guys a question, because we saw this at DevOps. Early days of infrastructure as code, it was, see yeah, go do that, go to the cloud. What the hell is the cloud? Yeah. So now we have this innovation formula where it's top down, bottoms up. Boardroom to dorm room. Boardroom, take that hill. We need AI and everything, infuse it, go. Boss, okay. And then the developer's just going nuts on open source so they're already self-forming. So you get the bottoms up going great, but the blocker from the top down is compliance, legal. You have, what seems like a blocker in that. How does that one get relieved? What happens next? I think the regulation is certainly something we all continue to keep an eye on to see where that goes. But there's so much. I mean, if you look at just how quickly organizations are innovating with the use of AI, both internally and for commercial technologies, I think it's going to become unstuck just from that sheer momentum. It's people are seeing the impact that it brings immediately to the users and to our customers and communities. So for me, I don't see the... You think it'll sort of stuff out very quickly? I think so. I don't think... Yeah, I would agree with that. We don't see anybody slowing down on that at all. There's also an interesting sweetener in that the compliance and the legal folks are prime beneficiaries of generative AI technology. You think about e-discovery and how transformative semantic search is there, or think about all the various regulatory compliance matrices that your typical international banker has to deal with, looking at product launches across different markets, cybersecurity standards and the like. So the fact that they can also pragmatically look at this technology and see their own lives transform kind of gives a new spin on this as well. This reminds me of the security business, how that industry was always a bunch of tools, now they're going to platforms. Now we don't hear a CUBE conversation ever. We hear words like, we build the insecurity from the beginning. Okay, finally, thank you. Now it's the data we're having that moment here to your point about compliance and governance. Do it from the beginning. The dividends are better. What are those dividends? What's the obvious? I mean, sales pitch or not sales pitch, but if you don't do it, that's what's going to look like. If you do it, this is what success looks like. What are you guys, because this is happening. This idea of getting in from the beginning with compliance, otherwise the innovation won't happen. Well, I mean, starting with our own business, we found that we really had to put generative AI front and center with the M365 stack with Teams, with the Power Platform, with Azure. We really had to make sure that every product in the Microsoft portfolio had this natural language interface. Because again, we had to service citizen developers, we had to service experienced developers, and everybody could benefit from this idea of then having that natural language interface to unlock all of the potential of that product. And in doing so, it kind of paved a way for us to give a blueprint to our partners to say, hey, we've went on this journey, let's go on it together. And if there's learnings or building blocks that we could then offer you, similar to what we learned internally, that can maybe unlock even more business value. Yeah, and I think one of the core tenets of this for us was we didn't rush into this, maybe as hastily as some others, right? We were very pragmatic about how we wanted to engage with Microsoft, and we engaged with a number of our top customers to understand their use cases and why. And I think we went, you know, you go slow to go fast, right? So, I mean, it's helped us really pick up momentum internally with some of the work that we showed this week. I mean, we're talking about weeks' worth of development work, right? As opposed to, you know, six months or a year. Because you had the foundation, that's right. And you're talking about what used to be all these boring, but important data quality, governance, compliance, and now we'll bolt those on. As an afterthought, now it's fundamental. You know, it's got to be designed. So they're actually really exciting spaces. And to your point, John, enable that growth. Yeah, I mean, I think, and the other factor is, is that with the benefit of, like, seeing Microsoft scale, at Azure, these guys' customers have workflows and data is the intellectual property. It's not an asset anymore, it's actually IP. It's competitive advantage. You've got to protect it, you've got to use it and harvest it, build value, harvest value, deliver value. That's very data-driven. It's very customized to these applications and verticals. Customers that we're talking to, all of them are trying to figure out how they can use their data to competitive advantage, both internally and then monetizing it externally. And I think, as these organizations have started baking it in, as you said, they're understanding they can get to decision faster, which, at the end of this, it's not just about the analytics and the data, it's about how quickly can I outpace my competitors. Thanks for coming on. Really appreciate it. Final question would be just share your thoughts on how this goes forward, the relationship with Microsoft, and the impact of your ecosystem, because a billion dollar investment in industry verticals and solutions means you're going to see specific applications taking advantage of domain expertise and data, the human, the pilot, and you got to scale it and you got to get more partners on board. So like, share your thoughts on it. I mean, it's complicated. The tip of the iceberg, right? I mean, there's a culminate with an announcement this week that is from all the engineering work that's been done, and there's more to talk about that at Innovate for us next year. So for us, this is the beginning of, and the scaling our industry solutions with Azure is going to be a critical component of that because there's the platform, parts of our business and the customers that we talk to, and then there's the industry solutions that they're eager to adopt and get integrated with AI. So, it's a- Final word. Yeah, the reason we're so excited about working with SAS is we both have this very industry aligned and industry centric view of how we're going to actually transform our mutual customers together. And in that, we've brought these engineering teams together to look at sort of the variety of emerging, Microsoft Cloud for industry components that we're developing, the work that SAS is doing across their entire flagship line of products like Workbench and then to bring that together and to actually take practical real-world scenarios, accelerate them and then learn as engineering teams together and then re-incorporate those learnings back into the products to make both teams better. I mean, what a future ahead. Using data to enable the enterprise for AI is a great opportunity for you guys. Congratulations. And thanks for sharing the data here on theCUBE. Thanks guys. I was bringing you all the data here on theCUBE from SAS Explorer. I'm John Furrier, Dave Vellante. We'll be right back with our next guest after this short break.