 We're back at GTC 2024 in San Jose. This is theCUBE and Scott Bills is here. He's the Vice President at Dell Professional Services. Scott, thanks for making some time. Yeah, thank you for having me. So you're welcome. I mean, services, we always say, rubber meets the road with services. That's where actually the business value is realized. It reduces risks for customers. But wow, we're talking about a whole new era here, right? And of course services, you go with the flow, right? If the customer says, hey, this is what we want. You guys figure out how to deliver it. You're obviously doing a lot of infrastructure. You've got a big ecosystem, but as I say, we're entering a new era now. So how are you guys, you know, first of all, what do you think of the show? And then let's get into how you guys are accommodating this AI world. Yeah, I love the show. Love the energy. Love the announcement we had with the Dell AI factory, which our services is gonna obviously help support from a Dell standpoint. Really what we're looking to do from a services point of view and how we're thinking about the space is that we want to help customers from day zero to day two plus on their AI transformation journey. So that starts with helping them understand what are the use cases? What's the strategy? How do you assess readiness? How do you prepare the organization for AI? Then how do you think about data readiness, prepare the data, engineer the data, ingest the data, scale that. How do you stand up the infrastructure, the software, the hardware, deploy the models, optimize, and then scale over time? So our goal with Dell and Dell Professional Services is to really help our customer through that end-to-end journey but also meet the customers where they're at. So if we do have a customer, for example, who has addressed the use cases, figured out the strategy, we can help them with the data or we can help them with performance optimization. It's meeting them where they're at and working with them with the AI factory to provide a real easy button for AI enterprises. Let's unpack that AI factory a little bit. So we saw Michael Dell at the keynote, right up front, he had a little bit of an entourage with him, which was awesome. And then Jensen gave an unbelievable shout-out. He said, something to be effective, there's no better company in the world than building end-to-end AI systems than Dell. It was like floor-drop moment. It was Michael Dell, he was ready to take the orders in a good way, good tongue-in-cheek joke, but he's ready to take the orders. So that was quite amazing. But what is the AI factory? What's that partnership like? How can we expect it to unfold? Yeah, it's really an end-to-end solution for enterprises to really provide them that easy button for AI. So it brings the breadth of Dell infrastructure and hardware, so compute, storage, networking, workstations, add devices, laptops, plus NVIDIA AI infrastructure and their software stacks, including the new MIM microservices software that they'll be bringing to market. Plus our combination of Dell and NVIDIA professional services as well to provide a real turnkey solution for enterprises to help them address their biggest challenges. They're most complex issues around adopting AI use cases and driving those into operations and production. Let's talk about what's happening in the field, what customers are asking you for. I mean, big picture. You know, tech spending is what it is. I mean, there's a macro headwind. Everybody talks about it on their earnings calls. Our data with our partner ETR shows that about 44% of customers tell us that they're funding AI, gen AI specifically, by taking from other budgets. Okay, so it's not just like, here's a checkbook, go. CFOs are being very careful. Two thirds of the customers expect their gen AI ROI inside of 12 months. Not to put pressure on you. So we know there's a lot of experimentation going on. What are you seeing in the field? How are you thinking about that, helping them get that ROI? What are you seeing? It's really that upfront process of identifying the use cases, prioritizing them, and really understanding where the business value is from AI. So it's understanding how do you apply it, how do you integrate it into your business processes, applications, and how do you make it real? So it's really that first step of understanding how AI drives value. As you're thinking about POCs, as you're thinking about use cases and driving that into production, it's having that upfront lens that assesses that upfront and prioritizes based on that. So it's got to be an interesting, I mean, you've, I'm sure, seen many waves. I don't think we've ever seen one like this, but we've seen waves. So we know what a wave looks like and feels like, and a lot of times it's like, hey, we got to do something in AI, or we got to do something in X, and that's sort of what's happening now. The boss says, hey, what's your AI strategy? I don't really have a good one. So what is that starting conversation like? You start with the data, you say you start with the use case, but it's like, what are they asking you? What are customers asking you? How do I drive more productivity? I'm sure they're asking, how is Dell using AI internally? How are others in our industry using it? So are you seeing any patterns that are emerging? A lot of the large enterprises are very, very similar to the experience that we're seeing with Dell, which is they have hundreds of use cases they've identified across the organization, and the challenge is really where to get started. Where is the biggest opportunity to bang for the buck? How do you think about clusters of common use cases? How do you think about the patterns? How do you think about architecture from a solution standpoint, and a data strategy to support that as well, aligned with those clusters and use cases, but it's really around figuring out where the best opportunity is, just because in most organizations they're finding hundreds of potential applications for AI. Yeah, so in gymnastics they have the ratings, right? And there's the execution, and then there's the degree of difficulty, maybe figure skating, kind of the same thing, right? So I would imagine there's a similar analogy here, is if you've got really good data, and it's clean data, and it's high fidelity, and it's sort of low risk data, well that's going to be easier, but it may not deliver the business value. So where are people starting? Are they starting to get some quick wins in singles just to get some muscle memory going, or are they going after those big productivity hits? Quick singles I would characterize it as, and they're also focusing more on internal use cases, so functional optimization, things like content generation, software development, sales, support, services, things like that, where it's a lower risk environment internally, before they start to pursue AI use cases that may be more externally customer focused or customer facing, where there may be some more execution risk, or the impact of execution risk might be higher. So software development will be coding, is this correct? So that's got to be a big use case. Yeah, absolutely. What are the early returns saying in terms of the productivity improvements that people are seeing with developers? Yeah, it's significant, so it definitely is worth the investment that we're seeing in terms of the AI tools, the solutions to go drive that, the use cases, so the early impacts for software development from what we've seen is strong, it's positive. And then, what's the state of data? I mean it's, on the one hand I feel like the big data era and Hadoop and all that stuff has sort of really got us thinking about how to get our data as states in order, but at the same time, so much more data was created with cloud and mobile and social, and we just became inundated, it was almost really quite difficult to keep up, and then sort of Hadoop didn't get us where we needed to go, so we sort of pivoted, started to develop pretty sophisticated data pipelines, which probably helped. A lot of that's focused on analytics. I still got the transactions over here, my analytics over there. I've got different data types structured and unstructured, so it's a real challenge, but what are you seeing in your customer base in terms of the state of data? I'm sure it's a wide spectrum, but are we in good shape or do we have to fix our data state before we can get that AI ROI? Fix that data, state is generally what we're seeing, and I think the key goes back to what we were saying before, which is to understand what your data strategy should be and how you should be managing data and leveraging data, you need to understand what the use cases are going to be, and I think most organizations, rather than figuring out what data should look like on a use case-by-UK basis, are going to have to develop a more comprehensive strategy for thinking about data across use cases and AI adoption more generally within the organization. Scott, how about the, we've put forth, you know, my gosh, a year ago now, the power law of gen AI, where you got some big, big models, big size of models, but then you've got a long tail, smaller models, very specific domain, domain specificity within, you know, let's say, I think great use case would be financial services or healthcare, virtually any industry actually, so what are you seeing in terms of that, sort of sovereign AI, private AI, sort of on-prem AI, if you will? A lot more experimentation with the domain-specific AI or LLMs, foundational models, versus organizations looking to build and train their own models. You know, I think when we were looking back six months ago, there was a sense that people would be building and training their own models internally, and what we're finding is that there's more gravitation to the smaller, more domain-specific, kind of smaller parameter models than, you know, developing models from scratch on their own and training models. Are people leaning into Retrieval Augmented Generation? Are they looking to you to help them develop that? Absolutely. And how are they going about it? Are they trying to do it out of the box? Are they kind of cobbling it together on their own? Well, that's one of the exciting things we announced this week as well, in addition to the AI factory, is the RAG solution that we have developed jointly with NVIDIA and Dell, Common Architecture, and then we do have a set of services around that as well to help customers rapidly operationalize, implement, and deploy their first set of RAG use cases using a common framework, Common Architecture. So what would that entail? Would I take, basically, a corpus of data, make sure it's clean, put it inside of a Dell object store and get the vector database piece of my choice? I mean, you may have opinions, but I presume you're saying to customers, I kind of know Dell, use whatever you want, you know? And if you want an opinion, we'll give you one. But, okay, so it's sort of open. I can choose my core database, my vector database, and go. And then, essentially, you start getting an LLM or a RAG out of the box, is that right? Yeah, that's correct. And then you start working it and she's human. Help them to deploy the vector database, the embeddings, figure out what the specific use case would be, what the model should be, and then, yeah, get them implemented and help them operationalize that and start to see the initial results and train that over time as well. What about the broader ecosystem? I mean, Dell's a product company, but you've got services that are accelerants. You also have relationships. Dell has relationships with GSIs. You have relationships with software companies. What's the ecosystem look like? How is that reshaping for AI? Yeah, we're developing the specific strategy for working with GSIs and their customers and how they're specifically around vertical use cases and leveraging their industry-specific expertise. We're also heavily relying on the ISV community to bring technical expertise, IP and leadership in specific use cases as well. So we're working with the broader partner community in a number of different ways, both from a go-to-market standpoint, but then also delivery standpoint as well. Dell's service is a secret weapon, Scott. Thanks so much for spending some time on theCUBE. Yeah, thank you very much. Enjoyed it. You're welcome. Okay, we're here covering the AI era, GTC 2024, Dave Vellante and John Furrier. We'll be right back right after this short break. You're watching theCUBE.