 From Times Square in New York City, it's theCUBE. Covering IBM's Change the Game, winning with AI. Brought to you by IBM. Hi everybody, we're back. My name is Dave Vellante and you're watching theCUBE, the leader in live tech coverage. We're here with Scott Hebner, who's the VP of marketing for IBM Analytics and AI. Scott, it's good to see you again. Thanks for coming back in theCUBE. It's always great to be here. I love doing these. So one of the things that we've been talking about for quite some time in theCUBE now, we've been following the whole big data movement since the early Hadoop days. And now AI is the big trend. And we always ask, is this old wine new bottle or is it something substantive? And the consensus is it's real innovation because of the data. What's your perspective? No, I do think it's another one of these major waves. And if you kind of go back through time, there's been a series of them, right? We went from sort of centralized computing into client server, right? And then we went from client server into the whole world of e-business and the internet back around 2000 timeframe or so. Then we, from internet computing into cloud, right? And I think the next major wave here is that next step is AI and machine learning and applying all this intelligent automation to the entire system. So I think it's not just a evolution. It's a pretty big change that's occurring here. And particularly the value that it can provide businesses is pretty profound. Well, it seems like that's the innovation engine for at least the next decade. It's not Moore's law anymore. It's applying machine intelligence and AI to the data and then being able to actually operationalize that at scale with the cloud-like model, whether it's on-prem or off-prem. Your thoughts on that? Yeah, I mean, I think that's right on because if you kind of think about what AI is gonna do and in the end, it's gonna be about just making much better decisions, evidence-based decisions, your ability to get to data that is previously unattainable, right? Because it can discover things in real time. So it's about decision-making and it's about fueling better and more intelligent business processes, right? But I think what's really driving sort of under the covers of that is this idea that our clients really getting what they need from their data. Because we all know that the data's just exploding in terms of growth. And what we know from our clients and from studies is only about 15% of business leaders believe that they're getting what they need from their data. Yet most businesses are sitting on about 80% of their data that's either inaccessible, unanalyzed or untrusted, right? So what they're asking themselves is how do we first unlock the value of all this data? And they knew they have to do it in new ways. And I think the new way starts to talk about cloud-native architectures, containerization, things of that nature, plus artificial intelligence. So I think what the market is starting to tell us is AI is the way to unlock the value of all this data and it's time to really do something significant with it. Otherwise, it's just going to be marginal progress over time. They need to make big progress. The data is plentiful, insights aren't. And part of your strategy has always been to bring insights out of that data and obviously focused on client outcomes. But a big part of your role is not only communicating IBM's analytics and AI strategy, but also helping shape that strategy. So how do you sort of summarize that strategy? Well, we talk about the ladder to AI because one thing, when you look at the actual clients that are ahead of the game here and the challenges that they face to get to the value of AI, what we've learned very, very clearly is that the hardest part of AI is actually making your data ready for AI. It's about the data. It's sort of this notion that there's no AI without an information architecture, right? You have to build that architecture to make your data ready because bad data will be paralyzing to AI. And actually, there was a great MIT Sloan study that they did earlier in the year that really dives into all these challenges. And I believe if I remember correctly, about 81% of them said that the number one challenge they had is their data. Is their data ready? Do they know what data to get to? And that's really where it all starts. So we have this notion of the ladder to AI at several very prescriptive steps that we believe through best practices you need to take to actually get to AI. And once you get to AI, then it becomes about how do you operationalize it in a way that it scales, that you have explainability, you have transparency, you have trust in what the model is. But it really much is a systematic approach here that we believe clients are gonna get there in a much faster way. So I have a picture of the ladder here. It starts with collect. That's kind of what we did with Hadoop. We collected a lot of data because it was inexpensive and then organizing it. It says create a trusted analytics foundation, still building that sort of framework and then analyze and actually start getting insights on demand and then automation. That seems to be the big theme now is how do I get automation? Whether it's through machine learning, infusing AI everywhere, be a blockchain is part of that automation obviously. And then ultimately getting to the outcome, you call it trust, achieving trust and transparency. That's the outcome that we want here, right? Yeah, I mean, I think it all really starts with making your data simple and accessible, which is about collecting the data and doing it in a way that you can tap into all types of data regardless of where it lives. So the days of trying to move data around all over the place or heavy duty replication and integration, let it sit where it is but be able to virtualize it and collect it and containerize it so it can be more accessible and usable. And that kind of goes to the point that 80% of the enterprise data is inaccessible, right? So it all starts first with, are you getting all the data collected appropriately and getting it into a way that you can use it? And then we start feeding things in like IoT data and sensors and it becomes real time data that you have to do this against, right? So notions of replicating and integrating and moving data around becomes not very practical. So that's step one. Step two is once you collect all the data doesn't necessarily mean you trust it, right? And so when we say trust we're talking about business ready data. Do people know what the data is? Are there business entities associated with it? Has it been cleansed, right? Has it been, you know, take out all the duplicate data? What do you do in a situation when data is, you know, you have different sources of data that are telling you different things? Like I think we've all like been on a treadmill where the phone, the watch and the treadmill actually tell you different distances and what's the truth, right? So the whole notion of organizing is getting it ready to be used by the business and applying the policies, the compliance and all the protections that you need for that data. Step three is the ability to build out all this ability to analyze it and to do it on scale, right? And to do it in a way that everyone can leverage the data. So not just the business analysts but you need to enable everyone through self service. And that's the advancements that we're getting in new analytics capabilities that make mere mortals able to get to that data and do their analysis. And if I could inject, I mean, the challenge with the sort of traditional decision support world is you had maybe two or three people that were like the data gods. And you had to go through them and they would, you know, get the analysis and that's just the agility wasn't there. So you're trying to democratizing that, putting it in the hands. Maybe the business user is not as much of an expert as the person who can build the cube but they can find new use cases and drive more value, right? Actually, from a developer that needs to get access and analytics and fuse into their applications to the other end of the spectrum which could be a marketing leader, a finance planner, someone who's planning budget, supply chain planner, right? So it's that whole spectrum, right? And not only allowing them to tap into and analyze the data and gain insights from it but allow them to customize how they do it and do it in a more self service. So that's the notion of scale on demand insights. It's really a cultural thing enabled through the technology. And with that foundation then you have the ability to start infuse where I think the real power starts to kick in here. So I mean, all that's kind of making your data ready for AI, right? Then you start to infuse machine learning everywhere, right? And that's when you start to build these models that are self learning that start to automate the ability to get to these insights into the data and uncover what has previously been unobtainable, and that's where the whole thing starts to become automated and more real time and more intelligent, right? And those models then allow you to do things you couldn't do before with the data that they're saying they're not getting access to. And then of course, once you get the models just because you have good models doesn't mean that they've been operationalized that they've been embedded in applications, embedded in business process, that you have trust and transparency and explainability of what is telling you. And that's that top tier of the ladder is really about embedding it, right? So that into your business process in a way that you trust it. So we have a systematic set of approaches to that best practices. And of course we have the portfolio that would help you step up that ladder. So the fat middle of this bell curve is kind of this maturity curve is kind of the organize and analyze phase. That's probably where most people are today. And what's the big challenge of getting up that ladder? Is it the algorithms? I mean, what is it? Well, I think it clearly with most movements like this starts with culture and skills, right? And the ability to just change the game within an organization. But putting that aside, I think what's really needed here is an information architecture that's based in the agility of a cloud native platform that gives you the productivity and truly allows you to leverage your data wherever it resides. So whether it's in the private cloud, the public cloud on premise, right? Dedicated, no matter where it sits, you wanna be able to tap into all that data. Because remember, the challenge with data is it's always changing not only in the sources, but the actual data. So you need an architecture that can handle all that. Once you stabilize that, then you can start to apply better analytics to it. And so yeah, I think you're right. That is sort of the bell curve here. And with that foundation, that's when the power of infusing machine learning and deep learning and neural networks, I mean, those kind of AI technologies and models into it all, this takes it to a whole new level. But you can't do those models until you have the bottom tiers under control. Right, setting a foundation, building that framework and then applying. What developers of AI applications, particularly those that have been successful, have told us pretty clearly is that building the actual algorithms is not necessarily the hard part. The hard part is making all the data ready for that. And in fact, I was reading a survey the other day of actual data scientists and AI developers and 60% of them said the thing they hate the most is all the data collection, data prep, because it's so hard. And so a big part of our strategy is just simplify that, make it simple and accessible so that you can really focus on what you wanna do and where the value is, which is building the algorithms and the models and getting those deployed. Big challenge and hugely important may IBM's 100 year old company is going through its own digital transformation. We've had Inderpal Bandarion talking about how to essentially put data at the core of the company. It's a real hard problem for a lot of companies who were not born five or seven years ago. And so putting data at that core and putting human expertise around it as opposed to maybe having whatever as the core, humans or the plant or the manufacturing facility, that's a big change for a lot of organizations. Now at the end of the day, IBM and IBM sell strategy, but the analytics group, you're in the software business. So what offerings do you have to help people get there? Well, in the collect step, it's essentially our hybrid data management portfolio. So think DB2, DB2 warehouse, DB2 event store, which is about IoT data. And that's where big data in Hadoop and all that with Hortonworks, that's where that all fits in. So building the ability to access all this data, virtualize it, do things like query plex, things of that nature is where that all sits. Query plex, meaning that to the data virtualization capability, to get to the data. Define a query and don't worry about where it resides. We'll figure that out for you kind of thought, right? In the organize, that is infosphere. So that's basically our unified governance and integration part of our portfolio. So again, that is collecting all this, taking the collected data and organizing it and making sure you're complying with it, whatever policies and just making it business ready, right? And so infosphere is where you should look to understand that portfolio better. When you get into scaling analytics on demand, that's Cognos analytics. It is our planning analytics portfolio. And that's essentially our business analytics part of all this. And some data science tools like SPSS for doing statistical analysis and SPSS modeler for doing statistical modeling, things of that nature, right? When you get into the automate and the ML everywhere, that's Watson studio, which is the integrated development environment, right? Not just for IBM Watson, but all, it has a huge array of open technologies in it like TensorFlow and Python and all those kind of things. So that's the development environment. Then Watson machine learning is the runtime that will allow you to run those models anywhere. So those are the two big pieces of that. And then from there, you'll see IBM building out more and more of what we already have, but we have Watson applications like Watson assistant, Watson discovery. We have a huge portfolio of Watson APIs for everything from tone to speech, things of that nature. And then the ability to infuse that all into the business processes is sort of where you're gonna see IBM heading in the future here. I love how you brought that home. And we talked about the ladder and it's more than just a PowerPoint slide. It actually is fundamental to your strategy. It maps with your offerings. So you can get the heads nodding with the customers. Where are you on the maturity curve? Here's how we can help with products and services. And then the other thing I'll mention that we kind of learned when we spoke to some others this week and we saw some of your announcements previously, the red hat component, which allows you to bring that cloud experience no matter where you are. You've got technologies to do that. Obviously, you know, red hat, you guys have been sort of birds of a feather and open source because your data is gonna live wherever it lives. Whether it's on-prem, whether it's in the cloud, whether it's in the edge and you wanna bring sort of a common model whether it's containers, Kubernetes, being able to bring that cloud experience to the data, your thoughts on that? Yeah, I mean, and this is where the big deal comes in is for each one of those tiers, so the DB2 family, InfoSphere, right, Business Analytics Cognos and all that and Watson Studio, you can get started and purchase those technologies and start to use them, right, as individual products or software as a service. What we're also doing, and this is the more important step into the future, is we're building all those capabilities into one integrated unified cloud platform that's called IBM Cloud Private for Data. Think of that as a unified collaborative team environment for AI and data science, completely built on a cloud-native architecture of containers and microservices that will support a multi-cloud environment. So, IBM Cloud, other clouds, you mentioned red hat with OpenShift. So over time, by adopting IBM Cloud Private for Data, you'll get those steps of the ladder all integrated into one unified environment. So you have the ability to buy the unified environment, get involved in that, and it's all integrated, no assembly required kind of thought, or you can assemble it by buying individual components or some combination of both. So a big part of the strategy is a great deal of flexibility on how you acquire these capabilities and deploy them in your enterprise. There's no one size fits all, we give you a lot of flexibility to do that. And that's a true hybrid vision. I don't have to have just IBM and IBM Cloud, you're recognizing other clouds out there, you're not exclusive, some companies, but that's really important. It's a multi-cloud strategy, really. It's a multi-cloud strategy. So that's exactly what we mean is we recognize that most businesses, there's very few that have standardized on only one cloud provider, right? Most of them have multiple clouds and then it breaks up a dedicated private public. And so our strategy is to enable this capability, think of it as a cloud data platform for AI across all these clouds, regardless of what you have. All right, Scott, thanks for taking us through the strategy. I've always loved talking to you because you're a clear thinker and you explain things really well and in simple terms, there's a lot of complexity here, but it's really important as the next wave sets up. So thanks very much for your time. Great, always great to be here. Good to see you. Thank you. All right, thanks for watching everybody. We are now going to bring it back to Cube NYC. So thanks for watching and we will see you in the afternoon. We've got the panel, the influencer panel that I'll be running with Peter Burris and John Furrier. So keep it right there, we're right back.