[00:00:03] Today, we're going to be tackling one of the biggest challenges in modern data management. How to unify, govern, and personalize enterprise data without moving it. And to guide me through this crucial topic, I'm honored to welcome my guest today. He's the CEO of Denodo, a company that has been at the forefront of logical data management, enabling organizations to seamlessly integrate, access, and govern their data across so many different silos.
[00:00:33] Without the need for duplication or migration. Now we're at this moment in time where AI is playing a bigger role than ever. With Denodo integrating retrieval augmented generation or RAG techniques for large language models and partnering with major players from Google Cloud and Amazon, Nvidia, and OpenAI to supercharge enterprise data and AI applications.
[00:00:56] That alone makes it incredibly exciting for me to get my guest on today. So we're going to discuss how Denodo's logical data management platform can help transform corporate data strategies, making AI more intelligent, and also ensuring that enterprises can access the right data at the right time, securely and efficiently.
[00:01:18] We'll even have a sneak peek at the upcoming O'Reilly book, logical data management, which takes a deeper dive into the approach. But how can businesses unlock the full potential of their data while also maintaining security, compliance, and governance? And what role will AI play in reshaping how we interact with information? Enough scene setting for me. Let's get today's guest on now.
[00:01:44] So a massive warm welcome to the show. Can you tell everyone listening a little about who you are and what you do? Well, my name is Angel Vina, the founder and chief executive officer of Denodo, as well as president of the board of directors of the company. I oversee the overall corporate strategy, operations, worldwide expansion of Denodo.
[00:02:07] Under my tenure, Denodo has grown into a global leader in data management with over 1,000 customers, 30 offices in 25 countries, and about 700 and something employees. My professional career was in academia before entrepreneurship with a strong academic foundation.
[00:02:30] I did my PhD in telecommunications engineering in the Technical University of Madrid and postdocs as a research visitor at UCLA in California, also at the Stanford University. And then I continue as professor in several universities in Spain.
[00:02:51] As chief executive officer of Denodo, I remain steadfast in a mission to revolutionize how business harnesses the power of data, empowering organizations with simplified data access to any data source they have in their data ecosystems and making data consumption as easy as possible.
[00:03:14] Trying to not to fit any type of business consuming application in an ever evolving digital landscape. And beyond my professional accomplishments, deeply committed to social responsibility in my communities, as I would say many citizens do,
[00:03:32] I really envision a world where technology not only drives business success, but also addresses societal challenges, emphasizing the importance of ethical and responsible use of data. Well, it's a pleasure to have you join me on the podcast. And every day I try and demystify a different area of technology or help business leaders understand the impacts it can have on their business.
[00:03:59] So to set the scene for today's conversation, Denodo's logical data management approach is something that intrigued me and something that I wanted to dig a little bit deeper on today. So can you tell me a little bit more about that logical data management approach and why you feel it's essential in today's data driven world, especially with AI now, of course, data is more important than ever? It is. It is. It is. It is. It is. There are many data challenges.
[00:04:26] It is very well known by everybody, the exponential volume, the diversity, the security fears, and the collecting costs around data management in corporations, in modern corporations today.
[00:04:40] They manage many applications and systems focused on different aspects of the business, from supply chain management to marketing automation, as well as a variety of data systems, such as data warehouses, data marts, data lakes, data lake houses.
[00:04:58] To be able to deliver data that meets all these requirements, organizations need to efficiently manage these complex data landscapes. And they have turned basically to two different approaches.
[00:05:16] One, which is a centralized approach where data is consolidated, is materialized in a warehouse, a data lake or a lake house, or logical data strategies, which take advantage of data lakes and anything that you may have to warehouse the data. But they are based on consolidated views of data across disparate data systems.
[00:05:37] This is very much our approach to the market, is what we propose to corporations to modernize their data ecosystem with additional capabilities to really have a logical base within their data ecosystems. And I'd love to find out a little bit more about the Donodo 9 platform, which I was reading about, and one of the reasons I invited you on the podcast today.
[00:06:02] So for people listening, hearing about it for the first time, can you tell me a little bit more about its features, what makes it different from other things out there, and why for many it's becoming one of the leading choices for data management? What is it that you guys do differently? When we say a logical approach, what we have is very unique capabilities in our technology to address the diversity of data systems that usually you find in modern corporations today.
[00:06:30] These unique capabilities, I mean, you can summarize in three areas. First is the ability to semantically unify data silos without the need of moving the data. That is a capability which is very important because most of the vendors today, when they talk about a semantic unification, moving data into a data lake, into a lake house. That is good, per se.
[00:06:59] But, I mean, the need is much wider, bigger than that. I mean, in a normal corporation, you have tens, in some even hundreds, in some others, even thousands of data systems that you need to unify. The ability to unify semantically, logically unify these data systems is very important. It's a capability that should be in all data stacks in a corporation. Second is the ability to federate data governance.
[00:07:29] Security, compliance are big topics today. The ability to fear is that you are secure, that you control the data feeds that you deliver for consumption to data applications. It's very important to control that with federated data governance capabilities. Most of the solutions that you find in the market today, they have centralized data governance.
[00:07:59] They govern a data system. What Denodo is proposing is to add governance capabilities in the logical layer, in that unification, that semantic layer that is on top of the data systems. Very unique to Denodo, especially to address modern architectural paradigms like data mesh, data fabric, that are very important for modernizing the data systems.
[00:08:23] And third, which is very important, it gets a lot of traction in the market for us, is the ability to hyper-personalize data consumption. And when we talk this extra layer or level of customization of data consumption, we refer very much about the ability to prepare data feeds, data sets, data products that are really ready for any particular audience.
[00:08:50] Understanding the management of the lifecycle of the data product beginning to end. This hyper-personalization at the right cost is also very unique with the abilities or the capabilities that we have to control and manage the consumption, the delivery layer of the data ecosystems. All that is very unique to Denodo.
[00:09:12] Add into that, these days with artificial intelligence and JGI into the picture, we also have very unique capabilities to support a rack augmentation in the development of JGI applications and the ability to interact with LLMs from SQL queries.
[00:09:36] Both areas of functionality in Denodo data platform are becoming very strategic, very important in our customer base to address the development of new JGI applications. And when I was doing a little research on you, one of the things that I noticed was Denodo's experience, significant growth over the years. There's got to be a story there. Is there anything you can tell me about that growth stories and some of the factors that might have contributed to its success?
[00:10:05] Because it's something that really seems to stand out from what I've been reading about you guys. Well, I think the one main factor is definitely this transformational nature of our technology. We radically changed the way organizations access and use data. And our vision was futuristic from the inception of the company.
[00:10:25] And that vision, based on the idea that data is organized in a distributed manner and requires semantic unification to facilitate data consumption, is more valid today than ever. For a reason, it's bigger volumes, bigger diversity, more variety of data consumers.
[00:10:44] Data is being democratized for consumption in different business units to different audiences to the extended user of the typical user of a data system. So all that is bringing more and more use cases and business-noted company like ours with our approach to the market. Growth also comes from a combination of focus, innovation to be the best at what we do.
[00:11:14] This focus has been with us from the very, very beginning. In addition to a focus on execution and operational excellence, deploying the right teams with the right skills to really support customers. And of course, a partner ecosystem that helps to address the market. We have an amazing total addressable market out there. We basically focus on companies with important data ecosystems with a significant number of data assets.
[00:11:43] And that focus is helping to really grow through our company today. I love that. And as this is a tech podcast, something we must talk about, of course, is AI or maybe generative AI. That seems to be where the buzz has been for the, what, coming up three years now, entering the third year. I'm curious, what role does all this play in your space?
[00:12:08] And are there any enhanced generative AI initiatives that you've implemented as a result of this trend? Because it's not just about you wanting to improve it. Your customers, I would expect, are coming to expect some of these technologies as well. Well, we bring the corporate data into the picture. And this is very important. We're not a company, a pure AI company doing LLMs or fine-tuning LLMs.
[00:12:36] But LLMs without the corporate data have very limited application for the enterprise. And what we do, unique, which other companies they don't do, is that we vectorize the metadata that describes the details and how to access the data in a corporate data ecosystem, in multiple systems at once.
[00:13:01] So we don't do LLMs on one single system, on one data lake. We do or we interact with LLMs and we augment the behavior of the LLMs with specific data coming from the structured data sources that are in a corporate data system.
[00:13:20] This augmentation that is based on the vectorization of the metadata that describes a data ecosystem is very unique to the point that we really bring the specifics of the structured data that matters to a corporation to a Gen AI application.
[00:13:37] Adding to that, and this is a little bit of the reverse, we have AI SDK that allows SQL queries to interact natively with LLMs. And this ability to interact with LLMs from SQL queries is also very unique to us in combination with the RAC augmentation. You join the two, I guess it's explosive.
[00:14:04] I mean, it's really bringing significant innovation to our customer base around artificial intelligence and Gen AI applications. And as I was looking through your social channels, something else that stood out to me was how you've recently announced a number of collaborations with some pretty big major players out there. And to name drop a few, Google Cloud, Vertex AI, Amazon Bedrock, NVIDIA, and OpenAI, which is just phenomenal. I mean, you've got all the usual suspects there.
[00:14:34] But can you tell me more about these partnerships, how they optimize the augmentation of LLMs, especially with corporate data through retrieval augmented generation or RAG as it's better known? Well, first, I mean, we need them. We are nobody without them, right? I mean, LLMs to really work closely with us. But definitely the ability to develop these partnerships is telling the market something, that we have something unique that matters to them as well.
[00:15:02] And we are working on a strategy to support all these major AI companies and cloud service frameworks for building generative AI applications. The recent announcements of integrations or partnerships with Google Cloud Vertex AI, Amazon Bedrock, LLMs, NVIDIA, NIME Inference Microservices, and OpenAI LLMs are part of this strategy.
[00:15:28] And all these frameworks offer developers access to foundational models and the tools to customize them for specific applications. Denoda augments the foundational models by adding access to corporate structure data that resides on any data source of an enterprise data ecosystem. And this is, as I mentioned before, a game changer, something that will create a total new wave of Gen AI business applications.
[00:15:57] And one of the things we've got to highlight is despite the promises of generative AI, many organizations are currently struggling to build applications that actually meet business goals and generate tangible value or measurable improvements. So how do you, at Denodo, help streamline the development of powerful Gen AI enterprise applications while also ensuring things like security, privacy, and responsible AI practices?
[00:16:24] Because it's quite a balance, and many are struggling to achieve this right now. Well, we launched a very successful Gen AI acceleration initiative every last year. We identified a number of customers that were open to explore the creation of new applications based on support for reaching LLMs and Gen AI applications.
[00:16:48] Companies like Perkins Coie, Sandia Labs in the U.S., NEC in Japan, Alexander Forbes in South Africa. And we tried not to bring a little bit of a flavor of our customer ecosystem, the PwC in Veltium, Festo in Germany. Most of these partners, if you want customer partners, because they are already customers of us, they are now moving into production. These first projects, there's some announcements that the company is going to make very soon.
[00:17:18] Some videos also showing these efforts. At the very end, using all these companies, using the Nodos Logical Semantic Model approach, we have been innovating using retrieval augmented generation techniques. And they developed the very first wave of use cases in their verticals, in their businesses, to really support different business units.
[00:17:56] And if we do look ahead, maybe even be able to do that in terms of the same thing in similar companies. We are very expected that everything is going to be one of the main producers of opportunities for us in 2025. And if we do look ahead, maybe even beyond 2025, are there any other key trends or indeed innovations that you foresee in data management and integration?
[00:18:22] Again, how are you at Denodo positioned to adapt and maybe even lead in these changes too? Well, there's a lot of change already going on. And just to adapt to the current wave of changes, this AI thing has been transformational. It has been reshaping completely our roadmap for the last, I can say, three, four years with the AI assistant copiloting, the experience of users of the platform. I mean, we're automating a lot of data processes.
[00:18:51] We have been adding new AI capabilities to our query-execured, foundational component of a logical layer, a semantic layer, across multiple systems, really manage query execution plans to really execute the queries in the physical systems. We have been adding a lot to really improve performance, reduce the cost, and provide observability and visibility and transparency to the way data is being used.
[00:19:19] What I was mentioning before about the support of integration with LLMs for J-AI applications, also very transformational, how that connects and links with our self-service interface in order to really elevate the experience of users with a data ecosystem that traditionally has been very heavy, very technical-based. These days, it's amazing.
[00:19:45] I mean, you get summaries, suggestions, I mean, automation of many query plans. You don't even need to be a SQL guy anymore. SQL gets generated automatically. So all this that is being part of the experience with modern data ecosystems in the most advanced, sophisticated customers is something that comes naturally with our platform users.
[00:20:09] Well, thank you so much for taking the time to sit down with me today and demystify this logical data management approach that you're leading with at Donodo. But before I let you go, I want you to leave one final gift for everybody listening. I have a Spotify playlist where guests can add a song to, and an Amazon wishlist where guests can add a book. I don't know why would you leave everyone with today, but what final gift would you like to leave everyone listening and why?
[00:20:36] Well, this has been a pretty serious podcast. That's in very deep technical topics, especially abstract topics. Yeah. So I would like to propose some upbeat music to relax the audience. from someone that I had the fortune to see this last summer in a concert live. It's the Carol G, the Colombian singer, with a song that is called Si Antes Te Hubiera Conocido.
[00:21:06] This is because if I were you before or something, I don't know what the exact translation would be of this, but I'm sure the listeners will love the rhythm of this nice bachata. Oh, beautiful. A classy addition to a Spotify playlist. I love that. I will add that straight away so people who listen can check that out. And of course, if people want to explore this logical data management approach a little bit deeper or just find out more information about Donodo or even contact you or your team
[00:21:35] if something in today's conversation resonated with them, where would you like to point everyone listening today? That's new. Our website, for sure. And there's something pretty new in our website, which is the first two chapters of the book that we're going to release this summer. The title of the book is Logical Data Management, of course. There's no other possible title for something coming from Donodo. I mean, they can enjoy the very first two chapters of this book already in our website today. Awesome.
[00:22:03] Well, I'll add links to everything there, including where they can get those first two chapters of the book. So more than anything, just thank you for bringing this topic to life today and your time and adding a touch of class to our Spotify playlist too. Thanks for your time today. Thank you. So a huge thank you to my guest today for shedding light on the future of data management in an AI era. I think it's clear that logical data management is no longer just a strategy.
[00:22:29] It is a necessity for organizations that are serious about navigating complex multi-cloud and hybrid environments. And Donodo's approach proves that data doesn't need to be moved to be useful, secure and AI ready. By federating governance, enhancing personalization and integrating AI with corporate data, businesses can ensure their data is not just accessible, but actionable. But of course, the big question that still remains is,
[00:22:59] are you and your organization ready to make that shift? How do you see AI and logical data management shaping the way that your business leverages information? Well, let's continue this conversation. Techblogwriteroutlook.com is my email address. LinkedIn, X, Instagram, just at Neil C. Hughes. Easiest guy in the world to find. Don't just hit follow. Follow me. Send a quick message. Let me know your thoughts.
[00:23:29] And remember, as a listener of this podcast, there is an exclusive preview to an early release preview of the Logical Data Management, the new book from O'Reilly there. I would urge you to check that out. I'll include a link and everything. But that's it for today. I will be back again tomorrow with another guest. Please join me again. We'll keep talking as long as you keep listening. I'll speak with you all bright and early tomorrow. Bye for now.
[00:23:59] Bye for now. Bye for now. Bye for now.