Cloudera EVOLVE 2025 | Marcela Vairo, IBM

Clips
More from Cloudera EVOLVE 2025

Marcela Vairo

VP Data & AI, Americas

IBM

play_circle_outline From Testing to Implementation: Ensuring Trustworthy Data Governance for Scalable AI Solutions

play_circle_outline Only 1% of enterprise data currently leveraged for AI applications, indicating significant growth potential.

play_circle_outline IBM's Watsonx platform integrates with Cloudera for enhanced AI model development and governance.

play_circle_outline Partnership between IBM and Cloudera supports tailored solutions for specific customer needs.

play_circle_outline Use cases in fraud detection, hyper-personalization, and HR transformation showcasing AI benefits.

Info
Transcript

Marcela Vairo, IBM

Marcela Vairo

VP Data & AI, Americas IBM

New York City energy meets IBM insight. TheCUBE sits down with Marcela Vairo, vice president of data and artificial intelligence for the Americas at IBM, as analyst Dave Vellante explores how hybrid and multi-cloud strategies meet the AI era at Cloudera EVOLVE25. Hear why “there is no AI without data,” and how IBM and Cloudera bring AI to where mission-critical data already runs.

Vairo details a maturing enterprise playbook. Only 1% of enterprise data fuels AI today, while more than 90% is unstructured and growing faster. Governance and data quality ... Read more

explore Keep Exploring

What has been the evolution of customer attitudes towards AI and data governance in recent years? add

What percentage of enterprise data is currently being utilized for AI, and what factors contribute to this low utilization rate? add

What details can be provided about the integration of Watsonx with Cloudera technology for AI model development? add

What is the process for integrating Cloudera data with IBM’s Watsonx and how do the companies collaborate with customers on AI solutions? add

What are some examples of how AI is being utilized across different industries? add

bolt Powered by CUBE AI

Marcela Vairo, IBM

search

Dave Vellante

>> Hi, everybody. Welcome to New York City. We're here in Hell's Kitchen. My name is Dave Vellante. We're kicking off Cloudera EVOLVE 2025. This is Cloudera's. They go to the customer. It's not a huge customer event, although there are a number of people here. The house is pretty packed here for the keynotes. Charles Sansbury kicked it off this morning to, as I say, a fairly packed house. Now, Cloudera is a company that got the big data era going and has evolved into the AI era. We heard a lot of messaging there that we're going to talk about throughout the day. So we're super excited to have Marcela Vairo here. She's the vice president of Data&AI at Americas for IBM. You've heard of them. Marcela, great to see you. Thanks for coming on.

>> It's great to see you. It's a pleasure being here today.

Dave Vellante

>> So your role, you spent a lot of time with customers in this area, the greatest financial center in the world. So talk about your role. What's your specific focus? Obviously data and AI, but what does that all mean?

>> So we are helping customers to be successful as AI-driven companies. What does it mean? It means preparing their data, getting their data architecture-ready, because there is no AI without data. This hasn't changed. So we are helping them to select the best use cases and to think about the return on investment, and how to grow and scale AI to transform the way they operate.

Dave Vellante

>> It's interesting, Rob Thomas, who said there's no AI without IA, talking about architecture, but since he came up with that tagline, things have evolved toward gen AI, agentic, and that fit perfectly into that narrative. You just sort of... IBM obviously has quite a heritage in AI and data, so it's plugged into that. My question is, how has that changed or evolved the conversation with customers that you're having?

>> I think, I would say, two, three years ago, most customers starting to test AI, gen AI, I would say, use cases. But now, as it evolves and as they start putting AI really into production, AI to work, they're realizing that if they don't have the correct data, if they don't trust their data, if they don't have governance, they won't be able to scale. I think this concept of data and AI coming up together, it's clear not only to the IT guys, I think this is a major change, but the line of business is also realizing that they need to figure out that data, they need to trust their data in able to scale with AI. So I think it's a much more mature conversation nowadays.

Dave Vellante

>> I think there's another part of that conversation as well that IBM has evolved. Again, prior to I like to call it the AI heard around the world with ChatGPT in November of 2022, I mean prior that, there was a recognition that while maybe we over-rotated on the public cloud a little bit, and certainly IBM's been hitting on the hybrid cloud narrative for quite some time. The Red Hat acquisition obviously accelerated that. But the narrative then and the value proposition was, hey, we can help you rebalance. So there's maybe some repatriation going on. Do your experimentation in the cloud, maybe go in production, or maybe even move some stuff back. But depending on the workload, doing things in a hybrid fashion or on-prem might make more economic sense. The narrative has changed and the value proposition has changed in the following way, and I'd love to get your thoughts. In addition to what I just said, it's become, well, we have production workloads on-prem and we're never going to move those into the cloud.

>> Completely agree.

Dave Vellante

>> They're hardened, they're the processes, they work. We've got the costs right. The skill sets and the business processes and the applications that run, they work and we're not going to mess with them. But we do want to add intelligence to them. Beyond deterministic machine learning, we actually want to apply generative AI, we want to have that natural language interface, et cetera. So the notion has expanded to bring AI to the data. And so I'm curious as to how prominent that is in the customer base and what that means in terms of how you're helping organizations.

>> Okay. So first of all, I completely agree with you, it's a hybrid world. So there will continue to be workloads that will stay on-prem and workloads that will start in the cloud or that will be in the cloud and not only one cloud. So it's a hybrid and multi-cloud world. Even though most companies have their cloud vendor of choice, their most important provider there, it will be a multi-cloud world and hybrid world. So to grow with AI, you have to figure out how to use your data because the differentiator would come by using your own data, the data that it's in your company and outside of that structure and unstructured data, that is where the intelligence for your AI applications and for your AI agents is coming from. So nowadays only 1% of enterprise data it's being used for AI. Think about the potential of that. And why is that? Again, because of governance, because of security people, because data being in silos. So once you figure it out how to scale and use and leverage your data that are spread into multi-cloud and hybrid environments, you are really going to be able to differentiate and to operate in a different way and ahead of your competition.

Dave Vellante

>> It's an interesting stat. I love when IBM throws out stats. You guys do hardcore research. And so I tend to trust IBM's data. Yes, it's marketing, but it's also well-researched because this is IBM's heritage. You like to back up what you say. And that's interesting, 1% of enterprise data for AI. And when I think about that, I'm like, well, wait a minute, what about machine learning and fraud detection? But a lot of that is transaction data, which is very small compared to the unstructured data that gen AI is now operating on. So I would think, first of all, the amount of data is exploding, but I would think increasingly that percentage is going to grow and customers .

>> I have another stat for you there.

Dave Vellante

>> Yeah. Give us the stats.

>> So more than 90% of the data is unstructured and it's growing three times faster than structured data. So it's really like the key, I would say the secret would be how do I leverage all this unstructured data in hybrid environments to fuel my AI agents and my AI applications? And by the way, another one, a final one for you, I promise, by I think 2028, at least 15% of the enterprise operations will be done completely autonomously by AI agents, agentic AI agents. So think about that.

Dave Vellante

>> By when?

>> By 2028.

Dave Vellante

>> Yeah, I mean, that's-

>> So again, you will operate in an autonomous way, at least 15. This is an average, so some companies will have more than that. You have to trust your data to be able to do that.

Dave Vellante

>> Yeah, you got to. So that brings me to governance and Watson AI for governance and also your relationship with Cloudera. So the strategies of the two companies are very much focused on hybrid and multi-cloud. Okay, check. I'm trying to understand your relationship and where the overlap is and where you pick up or they pick up where you leave off. Because of course we heard this morning in the keynotes heavy emphasis on lineage and governance. IBM obviously with Watson AI for governance. Ed data have a very, very heavy focus on governance. You always have. So explain the relationship, the partnership. It's lasted the better part of a decade now. Give us the details there.

>> So we have several customers from across industries such as the financial services segment, healthcare, public sector, industry, so et cetera, et cetera, that are running very important data on Cloudera technology. And we developed our Watsonx platform to help customers to leverage that data, so including Cloudera data, which is a core for those companies as well, to build and fill their AI models and to govern that as well. So Watsonx AI is completely integrated with Cloudera. If you want to quickly leverage Cloudera data to build your or to fine tune your models, your LLMs, you can use Watsonx AI for that. And once you have it all ready, of course you have to monitor and to govern your models and the results of that all the time. So that's Watsonx.gov. So they are completely integrated. Cloudera has fundamental data across several industries and enterprises, and we have a platform that help companies to leverage that data for AI and to govern that data for their AI models and applications.

Dave Vellante

>> So it's worth explaining. I mean if you go back to the end of the 2000s, 2008, 2009, that's sort of when Doug Cutting created Hadoop, it became this open source project. Prior to that, there was the data warehousing trend. You guys know that well. And what you would do, you would buy a bunch of expensive hardware, you'd shove all your data in there, you'd get a database license, and then that was your data temple. What Cloudera enabled was a dramatic reduction in the cost of being able to manage that data because prior to you would have a schema all built and it was very rigid. And so that era brought about no schema on write, you developed the schema as you read. What does that mean? It means it cut your cost dramatically and it gave you a lot more flexibility. But it also created things like data swamps. And so the whole data lake and the data lakehouse came out and became into vogue. Okay, great. So now we're in an era where we've got all this data, it's better organized than it was certainly two decades ago, but at the same time, now we're applying a probabilistic intelligence to that. And it's like, uh-oh, I need to take the next step in cleaning my data, governing my data because there's risks, et cetera. And everybody's really focused on the legal and privacy risks with gen AI. So what are you hearing from customers? Do you have examples where you and Cloudera have partnered up working with Watsonx AI or Watsonx.gov and Cloudera's lakehouse where you've helped customers sort of get to that next level?

>> So there are two topics that are becoming more and more talked between customers that are governance. How do I create a layer of governance to ensure that the users are using that data in the correct way? So this is one thing that happens all the time. And again, this partnership with Cloudera helps customer figure out how to access Cloudera data in a governed way with that governance layer. And not only Cloudera data because again, it's an hybrid world and the data's not sitting just in one, but there is very important core data there that they need to figure out. Another thing, so one major, it's a major thing, it's governance. And the second thing is data quality at the same time. And they are connected at the same time. So you need to figure out that you have a catalog of data that it's been accessed, and at the same time, who is accessing that data, it's correct, and you have to also have data security at the same time. So our platform guarantees all that and it's connected to Cloudera platform. So once again, you want to do more and to extract more and to scale with AI. And to scale, you have to put some, I would say, you have to put autonomous things to do that. We are putting it all together. That's why we built a corporation, I would say, a partnership there. It's not only a commercial partnership, it's a product partnership at the same time.

Dave Vellante

>> Okay, so there's some degree of automation between the two platforms.

>> That's the word. Yeah. That's the word. Yeah.

Dave Vellante

>> As a customer, I don't have to build scaffolding around it.

>> You won't be able to scale if you don't put automation into that. So you have to have the correct data architecture, but you also have to have automation in your AI and data tools to be able to scale with that. Because again, it's your operation that it's going to be transformed by AI.

Dave Vellante

>> How did that come about, Marcela? Was it a sort of eight years ago you guys started working in the field? That's how a lot of these things happen. You get joint customers and say, hey, maybe we should make it easier for them, build a solution, put in that automation, and then you get the product guys, and of course they're busy, so you got to get their attention. But so is that right? It's essentially, is it a solution? Is it a skew? Is it it just works because you find that you've got a Cloudera installation and IBM is in the account or can be brought into the account and that integration is made facile because of that product integration?

>> Exactly. So it's both. The products are integrated. So you can, for example, easily integrate Cloudera data into Watsonx.AI or into Watsonx.Governance for example. So the products, the platform is integrated. But more than that, again, each customer has its own data architecture, its own objectives and goals in terms of AI and agentic AI. So normally we would sit together, Cloudera, IBM team, and the customer, and build a solution for them. And of course, again, the partnership and the integrated products help to speed it up. But we'd normally sit together. It's not like, oh, I have already solution for you, it will solve all your problems. Normally we have to, when we are talking about the enterprise world, we have to sit together and figure out what's the best solution and the best architecture for that specific customer. And we are doing that again across several industries, financial sector, banking, healthcare, public sector, et cetera.

Dave Vellante

>> Okay. So there's a consultative aspect of what you bring. Of course IBM consulting is world-class. But every installation is not a snowflake, it's not custom because you've done that integration. So that lowers the cost, the upfront cost of planning and implementation. Are there particular... I mean you're very strong in financial services, obviously healthcare, regulated industries, all industries really. But we're here in New York. Are there particular benefits or patterns that you're seeing within certain industries in terms of what Cloudera and IBM are bringing in terms of customer value? We heard today, actually it was really interesting. I thought the keynotes, I don't know if you had a chance to hear them, were very meaty. I thought the AbbVie, which was Brian from AbbVie, which is biopharma, he said, "We look at value in terms of operational, analytical, and experiential value." So I thought that was pretty interesting. And he said they've saved 250,000 employee hours and all kinds of stats there. And they didn't fire people. They re-pointed them at strategic things. Are you seeing it happens like that?

>> Yes, we see a lot of cases, use cases for customer care. We see a lot of cases for fraud detection, a lot of cases of a potential for hyper personalization and creating new offerings for our customers. At IBM, as you say, you said about savings, and we at IBM, we have a very nice story. We drink our own champagne. So we were able to produce more than 3.5 billion in free cash flow just by using AI in several of our business areas. So we transformed the way our HR area operates, we transformed the way our procurement area operates, we are transforming sales. So it's across the board. And when you think about it, you really think about, again, we mentioned and we are in New York of course. Financial services is very strong here. In our partnership, we have several use cases, IBM and Cloudera in the financial services area. But it's a really cross-industry play, AI, when we think about it. And again, think about HR. Every large company has an HR department that can be completely transformed through the use of AI, for example.

Dave Vellante

>> We did a program at One Madison with the IBM consulting team about that HR transformation and people were raving about it. So that's pretty exciting. Where do you want to take this partnership from your perspective as a go-to-market pro? What's next with these initiatives?

>> I think if we think about where we are in the AI area, honestly, I think we are starting to scale that, especially with agenetic AI. So I see us doing more and more helping customers to leverage their data to agenetic AI because that is the thing that will transform their operations, not only from a cost savings perspective, this is the first step. And by the way, this is where most of companies will get the funding to invest more and more in gen AI. But I really see that's the first phase. The second phase and the differentiator will come when you use AI to do different things, to transform, not only to put automation in some of the workload that you have, but really to do it differently. Not only to eliminate some steps of your process, but really to produce new products and new services. So I think we are approaching an era of innovation as well with AI, and that's where the differentiator and that's where the winners will be. And I see us together on that partnership.

Dave Vellante

>> Well, picking up on a couple of interesting points. You mentioned earlier a stat, and I love the stats. 15% will be automated agents by, I think it was 2028.

>> Yeah.

Dave Vellante

>> And I've been waiting. You saw the MIT study, everybody talking about it, 95% failure rate on POCs. I don't know if it's that high. Many experiments are supposed to fail. So there's I think some good learnings that come out of that, but that was a good study. But the point I'm making is we've been waiting for AI to be self-funding because what's been happening in the early days of gen AI is we were stealing from other budgets. And so those other budgets were... And we're still seeing some macro pressures and uncertainty and all the public policy is uncertain. But a big part of it a couple years ago was that shifting of the budget, waiting for that AI to become self-funding. So there's going to be a crossover point to what you just said, and that's really where you start to hit the steep part of the S-curve I think. So anything else you want to share about what your organization-

>> Well, just to this point, I think something key is just not just playing with the technology. Oh, can you do this or that? Of course. Again, the technology there is there, it will work. The thing is what do you expect when you're doing a pilot? What's the outcome, the expected outcome? How do you scale that? If it works in this department, is it something scalable? If it works with that set data set, do you have all the data that you would need to scale? So if you plan for that, I think you're going to be ready to scale with AI. If you just do pilots just for test things just to say that you are doing something with AI... So it's not do something with AI, it's transform your company into an AI-driven company. It's a complete different perspective. So you have to start small but think big and plan on how to scale that in terms of processes, in terms of people, and of course in terms of your data architecture.

Dave Vellante

>> And have a framework to measure that value to the business and then that becomes a self-fulfilling prophecy because you're reinvesting. Marcela, thanks so much for coming on theCUBE. Really appreciate it.

>> Thank you. It was a pleasure being here today.

Dave Vellante

>> You bet. All right. And thank you for watching. This is Dave Vellante for the Cube at Cloudera EVOLVE 2025 from New York City. We'll be right back right after this short break.

>> Thank you.