Teradata Possible 2024 | Dan Spurling, Teradata

Clips
News
More from Teradata Possible 2024

Dan Spurling

SVP, Product Management

Teradata

Navigating the open and closed ecosystem: Enhancing AI integration and data democratization

The tech landscape is rapidly evolving. Companies increasingly apply AI integration, data democratization and open table formats to enhance decision-making and operational efficiency. As businesses strive to simplify data access and break down silos, the demand for robust, flexible systems that can seamlessly connect diverse data sources is more critical than ever.“I spend a lot of time with customers talking about, ‘What are you trying to do?’” said Dan Spurling, senior vice president of product management at Teradata Corp. “We’re seeing this huge push. There’s obviously the push around AI and the hype … that generative AI has driven

play_circle_outline Increasing Business ROI Through AI: Teradata's Approach to Open and Connected Ecosystem Integration

play_circle_outline Focus on meeting changing customer personas and AI use cases

play_circle_outline Future outlook on advancements in AI, generative AI, and machine learning for business solutions

play_circle_outline Teradata's investment in AI capabilities

Info
Transcript

Dan Spurling, Teradata

Dan Spurling

SVP, Product Management Teradata

Dan, a veteran at Teradata, emphasizes the importance of open table formats in the tech industry. Teradata focuses on providing flexibility and choice to customers with their powerful engine and workload management capabilities. Customer feedback is vital in developing their products. They offer solutions for data prep, model creation, and model ops, helping customers move from ideation to production. Teradata sees customers using AI and machine learning to tackle business challenges and provide real-time insights. Their strategy involves being a cloud data p... Read more

explore Keep Exploring

What makes Teradata's differentiated value and where does it truly differentiate in the market? add

What are some of the changes in personas that have been observed at the Teradata conference compared to a decade ago? add

What types of models and machine learning techniques are being used in addition to generative AI to address business challenges and meet customers using natural language? add

What are some of the AI capabilities that Teradata has been investing in for almost a decade and a half? add

bolt Powered by CUBE AI

Dan Spurling, Teradata

search

Savannah Peterson

>> Good afternoon Teradata community, and welcome back to Los Angeles, California. We are coming to the end of our power-packed day of coverage. My name's Savannah Peterson. Delighted to be joined by Rob Strechay. We've just been knocking out the super interesting and inspiring conversations.

Rob Strechay

>> I think it is... When you get to go really deep into the tech and understand how it actually translates to the customer outcome and use case, that's gold for me.

Savannah Peterson

>> We're making it real. One of the people to tell us how to make it real is Dan. Dan, you're a veteran VIP here on theCUBE. Thank you so much for being back on the show and joining us.

>> Absolutely, Savannah. Thank you. It's great to be here. Always a fun time.

Savannah Peterson

>> How is the week going for you? It's got to be so fun to see your community.

>> It is. First of all, it's fantastic to see so many people that I don't get to see very often from Teradata, so spending time with the team, spending time with our partners. A lot of partnerships go into actually being able to deliver what we have with the customers. And then, of course, being able to see so many different customers. I was just in London a couple of weeks ago with our customers over kind of that EMEA and APAC region. And then getting to come back home and seeing all our customers here in the US, it has been an amazing last couple of weeks and last couple of days.

Savannah Peterson

>> I can tell. Your energy is radiating. You're giving us that nice afternoon-

Rob Strechay

>> Yes....

Savannah Peterson

>> boost that we need here.

Rob Strechay

>> Yes. That's the coffee. That's just the coffee.

Savannah Peterson

>> I don't know about that. I don't know about that. There's only so much that can do.

Rob Strechay

>> Although the coffee game here is outstanding.

>> Yeah.

Rob Strechay

>> I give Teradata props for that, so I will... Anyways, but one of the things that, beyond the coffee, that gets me excited about this is really, and we talk about it a lot on theCUBE and have been because we see it as being key to where AI is going to go, is really open table formats and things of that nature. That has to be a really common discussion with your customers about, "Hey, we don't want to have to move all the data around. We're cool with you being the compute engine and really helping us format that, but help us with the open table format." How does that really translate to the product?

Savannah Peterson

>> We're going to talk about a lot of pieces around this because this is an area that I'm pretty passionate about. And I joined Teradata leading the engineering group, so my tendency is to start going deep on speeds and feeds and be like, "You know what the difference between Iceberg and what's going on with Delta? There's a Hudi on this."

Rob Strechay

>> Oh yeah.

>> I will channel as much as I can.

Savannah Peterson

>> You can get as nerdy as you want.

Rob Strechay

>> The community knows what Delta and Hudi and-

Savannah Peterson

>> Yeah, exactly. Yeah....

Rob Strechay

>> Iceberg are, so feel free to geek out.

>> I'm going to channel the why.

Rob Strechay

>> Yes.

>> I'm going to try to channel-

Rob Strechay

>> Yes....

>> the reason.

Rob Strechay

>> Yes.

>> So first of all, starting off, when we think about open table formats, stepping back from the tech, I spend a lot of time with customers talking about, what are you trying to do? And we're seeing this huge push. There's obviously the push around AI and the hype or especially that generative AI has driven,-

Savannah Peterson

>> Right....

>> but we're seeing this separate push around desire to drive massive amounts of data democratization inside their company, and big words that only mean like, hey, we should use more data when we make decisions, and we want to figure out how to get that data to the end users. So there's this natural evolution, and we're seeing it start to evolve in data products, where people are saying, "I want to be able to give trusted data products to everybody in my business unit. I want my marketing team, I want my sales team, I want my services teams to have access to trusted data, to be able to do things with it." And then they start running into these roadblocks of, well, that data is in 19 different systems. It's not like-

Savannah Peterson

>> Right....

>> the sales team's data is in one perfectly pristine system. It's spread across a bunch of different systems. And so we start seeing this evolution of people saying, "Oh, we need to do a data mesh, a data virtualization. We need to do all this stuff, and we're going to build this massive amount of complexity to give these people a data product." And then there's this push around open table format saying, "Do you really need to do that, or do you need to figure out how to have some type of trusted way to centralize this data and bring the engines or the tools to the data?" And so we're seeing open table formats emerge as a natural solution to this challenge of data silos, and then giving customers choice, giving them flexibility, giving them options. And it's making us, on the provider side, truly compete, not on just whether or not you're stuck in isolated environment, but do you really have the best engine? Do you really have the performance? And so we are confident in our engines, so we're like, "Bring it."

Rob Strechay

>> Yes.

>> And so we're jumping in, and we kind of put this under our idea of being truly open and connected.

Rob Strechay

>> Yeah. I think one of the things, we recently actually queried 100 customers that were both Snowflake and Databricks to ask them this specific question where, what are you using? What other DBMSs are you using? And things like that. And it was all over the map. They're not using just one thing. They're not even using just one cloud database.

>> Right.

Rob Strechay

>> I think one of the interesting things in there, in the open data formats, 70% said they're not there yet, but they're going towards an open data table format. Are you seeing that, to your point about being connected? And they just want things to work out of the box as well. And is that really where you're going to help them simplify this? Is that-

>> Yes. And it's fascinating, probably different from maybe the Snowflake or Databricks customers. Teradata customers are the largest of the large. And so when you were working with these customers, you would expect them to say they have the engineering capability or the technical capability to make things work. And so this idea of it working out of the box, you might think that, well, the really big companies don't need that. It's actually even more important for these really large companies. Because there is so much more complexity in their environment, they don't have the time to be specialty building things every single area. So we hear... Every conversation we hear, that we talk to is, what are you doing in the open table format space? And so every single conversation. They might not be there yet with production workloads. There's a whole lot of problems with cash invalidation, with multi-writers. There are legit issues that we, the industry, are working through, but that doesn't mean that it's not being a driving force in every architecture decision that we have. Yeah.

Savannah Peterson

>> That's interesting to hear, and I'm not surprised. You mentioned that this is a part of your overall open and connected strategy, philosophy. Can you give us the full bird's eye view of that, in general?

Rob Strechay

>> Yeah. Yeah. Open and connected. About when our CEO Steve took over, his push was around, how do we be an amazing cloud data platform for the masses? We didn't walk away from our on-prem capabilities. We were rock solid already there. We needed to be more solid in this modern cloud system. So we said we are going to focus on that. For us to focus on that though, we can't try to be all things to all people, and the reality is we shouldn't be. People should not come to Teradata and say, "I want perfect visualizations of all my data." That's not Teradata's differentiated value. Exact same conversation, "I want Teradata to do all of my workflow for ETL and ELT," probably not Teradata's strength. And then in the open table format space, "I want Teradata to be the sole owner of all of my data," probably also not our strength. And so we said, "Where are we going to truly differentiate?" And we're going to differentiate with our amazingly powerful engine and our workload management capabilities, our massively parallel optimization. And then everywhere else, we're going to try to leverage the best of breed. We're going to try to say, "What are the best tools out there that end users are using? What are the best tools for data movement? What are the best tools for data storage and data placement?" And really try to build an open and connected ecosystem of use the tool you love to use, get all the power of and the efficiency of Teradata's engine.

Rob Strechay

>> Yeah. I think one of the things you just hit on kind of, and I just to expand on it, was that it's really about where the customer is and meeting them where they are because the personas are changing of who's actually working with the data. AI's been around for more than a moment. It used to be ML and all of that. How do you really see, when you're designing these features and these products and the platform, to really meet those different personas and meet the customer where they are at?

Rob Strechay

>> Yeah, I think your point around the personas, you just have to look around the Teradata conference here and see the difference in personas than maybe you would've seen at a Teradata conference a decade ago, right? The data warehouse people that people are focused on, the underlying data, tools on data, move on data protection, those team members are still here. But you are seeing many, many customers who are thinking about giving data products out to customers, their customers who are internal to their business. And then you're seeing absolutely those AI or machine learning focused data scientists, citizen data scientists, data engineers that are focused on AI and ML workloads. You're seeing them, and they're saying, "Hey, I need access to not just the pristine, gold trusted data that's in Teradata. I also need access to all of this data." If you're a financial institution and you want to access to the raw ATM data, you're someone in the healthcare space and you want access to the raw population health data, you're someone in the transportation space and you want to access to the raw data coming off of FAA, there's so much additional data that doesn't need to be in the official corporate data warehouse, but a data scientist wants access to. And so for Teradata, we're so excited about being able to unlock that data and say, "That really trusted gold data, you want to build an AI solution that you are confident in? Absolutely, we've got that data. You want to do a lot of experimentation though and think about what else you could do, and you want access to all this other data that's spread out throughout your environment in open table formats or other open file formats? We got you too." And so being able, to your point... I think that the point you're making around personas, that persona is shifting, and we are making sure that we are there to meet that persona where they are with the tools, also with the data.

Savannah Peterson

>> I can imagine as you're building this out, customer feedback is critical for developing your next product. What's that feedback loop like for you and your team internally?

Rob Strechay

>> Yeah, it is been fantastic. Ever since I left, I didn't leave, ever since I jumped from the engineering side to really focus on product management, that feedback loop has been so crucial for us. And the feedback loops, plural, look like our executive advisory board sessions we have. They look like the customer feedback obviously that we have. It looks like a lot of competitive analysis that we do. It looks like feedback from industry analysts and the other people out there. And then finally, it is those people who are excited about some type of future that lock arms with us when we co-develop new things where we actually take all that feedback, and then we actually start really putting the rubber to the road and build something together. It's pretty phenomenal.

Rob Strechay

>> So to that extent, rubber meaning the road, and I think that we've been talking a lot about this because it came up in the keynote, and of getting value out of it and solving real-world business critical or business ROI generating things with AI, what are you seeing from your customers and the use cases that they're leaning into? And how is Teradata really enabling that?

Rob Strechay

>> I think that... I'm going to answer that question directly in a second, but I think stepping back, we are still seeing... And the industry, this is not a Teradata thing, the industry is still seeing so many people who have amazing ideas that stop at ideas. And so the idea for us of saying, idea double use, that the point for us is, how do you take that idea and do something with it? And so Teradata is spending a ton of time to say, how do we help with data prep? How do we make sure you've got data that you can trust when you start building? How do we help with model creation and multiple model creation, and learn ? And then how do we help you move from generally what is one persona, tied to your point earlier, to a different persona who takes that thing into production, the developer who builds an application around that model. And then once it's in production, we know that you can't just let it go. It needs to continually be refreshed. You might have feature drift, you might have model change. So being able to manage that in a model ops kind of way, we need to make sure that we, and this is the royal we, open and connected, but the industry is giving in almost like that beautiful workflow, and Teradata has to make that seamless to be able to truly get from ideation to production. We're seeing customers who are doing this already. Steve talked about some of the work we did. I'm also a giant Formula One fan. Same talk about some of the work we did and things that people might not have thought about Teradata with unstructured video data, and then being able to dynamically pull that in and then give information for strategy decisions in real time. We have another thing that we are launching with another partner who's here edge consulting around a complaint analyzer where you're taking in massive amounts of data from very disparate sources, some of them structured, most of them not, and then being able to then go back in real time and give prompt, and possibly even in natural language, prompt advice to a call center representative, to a sales representative, et cetera, or to someone making larger strategy decisions to say, "This is what the feedback is about your product in both a summarized way, but also in a human explainable way to help make decisions, either real time next best offer, next best action decisions, or of course, larger strategy. Where should we invest and put more investment?" So whether it's something as awesome as Formula One or something as critical as complaint analysis, lots of use cases. We're seeing generative AI already be delivered against open table formats, right? And I think it's important... I know I'm going a little long here, but I think it's important to highlight-

Savannah Peterson

>> No, great.

>> It's not just generative AI. Generative AI has the hype, but there are many models and machine learning that we have lots of experience with that holistically make a solution both address the business challenge, and then also meet that end customer where they are with natural language. So it's a super, super evolving space, and we are seeing customers, Teradata customers, as they're working to say, "How do I take this idea to production, and how do I really solve a business problem?" We're seeing it. We're right on the cusp of seeing those all start launching as we make that journey easier.

Savannah Peterson

>> If we could solve all the complaints at the speed of a race car, I think we would all be very, very happy. We've been enjoying the-

Rob Strechay

>> Although if you listen to the commentary, I think the various F1 drivers still have all their complaints. I don't know. I don't know if we can solve all of their complaints.

Rob Strechay

>> Well, I do like hearing their thought process about that. But like you said, I think having that real time data and being... It's a real edge type of use case where it's

Savannah Peterson

>> The ultimate edge....

Rob Strechay

>> the inference and being able to do that at scale-

>> Yes....

Rob Strechay

>> very quickly. And it would seem that... So your target customers are not just doing things in these massive things. They have to bring it to that edge, and that has to be another kind of, I guess you could say, little quirk to how you have to figure things out and when you're building out the product.

Savannah Peterson

>> Yes, absolutely true. And being able to think about edge, we don't have the amount of computing on the edge. So being able to think about intelligent decisioning on the edge, right? Pull back the decision, pull back the insights that need to be pulled back, but you do need to make some for real time action. You need to make some decisions-

Savannah Peterson

>> Absolutely....

>> on the edge. So Teradata, then this is a little bit separate, when we think about open and connected, we're also thinking about open and connected from a processor perspective, so we can have very low power compute on the edge that we're still driving analytic decisions, and then bringing back the really large scale analytic decisions back to massive compute that is maybe back in some type of data center or in the cloud providers. Yeah. And I cut you off, I'm sorry, earlier.

Savannah Peterson

>> Hey, it's your show. I'm just here to facilitate. Don't even worry about it. I'm curious, you talked a lot about we're right on the precipice here of seeing some of this actualized and realized at scale. On that vein, what do you hope to be able to say this time next year when you're back on the show that you can't say today?

Savannah Peterson

>> Yeah, that's a great question. Some of it, I can't say today. No, but-

Savannah Peterson

>> But this was our moment.

>> It was the moment. I could've... I could've-

Savannah Peterson

>> You're supposed to give us the scoop, Dan. Geez. Secret, secret. Yeah.

>> First of all,-

Savannah Peterson

>> The mic's not on. It's just a-

Rob Strechay

>> Yeah, it's just us.

Savannah Peterson

>> Yeah.

>> I think, first of all, when you step back from, why are we on the edge, I think that we are going to see a continued, unfortunately, a continued drop on the other edge of the high cycle, but I believe that that drop on the edge of the high cycle for generative AI will be a good thing. It will move away from people thinking that generative AI can do everything, which it might be able to do in the future, but not today, but generative AI will be able to do everything. And so no shade thrown at executive leaders, but when people like me say, "You know what? I want, team? I would like to have all of our software development be done by generative AI." Right? And we're not quite there yet. But if I was realistic about what we could do with test analysis, with story creation, with initial software creation, et cetera, then we can actually have clear success. So first, your point around why do I think we're on the precipice, I think we're on the precipice because I think we're going to be more realistic about what could we do today, and then start thinking and experimenting with what we do in two to three years. So first of all, that's what I think is going to drive that. And then what I'm excited about for Teradata is so many people look at Teradata and say, "Amazing data warehouse, amazing data platform," and they don't know that Teradata has been investing for almost a decade and a half on AI capabilities, machine learning, and then now AI, and then now generative AI capabilities. And with our announcement, for example, today around our partnership with Nvidia-

Savannah Peterson

>> Casual.

Savannah Peterson

>> Casual, just this little company, Nvidia. With our announcement there, with our announcement of launching our AI on-prems with GPU capabilities, that is seamlessly integrated with the cloud as well. So we talked about open and connected. You have a user who's used to using SageMaker, used to using something with Vertex, used to using Azure ML or OpenAI. They can work in those environments, and then push that down into production on-prem if they still have a lot of their data. We still have customers who have heavy regulatory requirements to stay on-prem, but that doesn't mean that those users that are building stuff can't work in a kind of really agile, nimble environment. Imagine being able to unlock those customers are like, "I'm stuck on-prem." Yeah, but that doesn't mean you're not going to innovate. We can help you innovate, and then be able to continue to meet those regulatory requirements. So what I'm excited about succinctly is with these announcements, I'm excited to see next year where people have realistic expectations and they get past these barriers to launching some generative AI, but true AI apps and being able to solve business problems.

Savannah Peterson

>> I love it, Dan. I can't wait till we have the chance to talk about it with you here back on the stage. Thank you so much for taking the time.

Rob Strechay

>> Thank you so much. It was great to talk with both of you, as always.

Savannah Peterson

>> Yeah. Yes. And Rob, always a pleasure, man. We are almost there.

Rob Strechay

>> Yes.

Savannah Peterson

>> Two more left. Wow, it's crazy. I hope you're all enjoying this as much as we are, wherever you may be tuning in. We're in Los Angeles, California, here at Teradata Possible. My name's Savannah Peterson. You're watching theCUBE, the leading source for enterprise tech news.