 Live from Miami, Florida, it's theCUBE. Covering IBM's data and AI forum. Brought to you by IBM. Welcome back to the port of Miami everybody. This is theCUBE, the leader in live tech coverage. We're here covering the IBM AI and data form. Of course the centerpiece of IBM's AI platform is Watson. Beth Smith is here, she's the GM of IBM Watson. Beth, good to see you again. You too, always good to be with theCUBE. So, awesome, love it. So give us the update on Watson. We know it's beyond jeopardy. Oh wow. Yeah, that was a long time ago now. Right, but that's what a lot of people think of when they think of Watson. How should we think about Watson today? So, first of all, focus Watson on it being ready for business. And then a lot of people ask me, so what is it? And I often describe it as a set of tools to help you do your own AI and ML. A set of applications that are AI applications where we have pre-built it for you around a use case. And then there's examples where it gets embedded in a different application or system that may have existed already. In all of those cases, Watson is here tuned to business, enterprise, how to help people operationalize AI so that they can get the full benefit because at the end of the day it's about those business outcomes. Okay, so the tools are for the super geeks who actually want to go in and build their own AI. That's right. The apps are, okay, it's pre-built, right? Go ahead and apply it. And then the embedded is, we don't even know we're using it, right? That's right, or you may. Like Q-Radar with Watson has an example of using Watson inside of it or open pages with Watson. So sometimes you know you're using it, sometimes you don't. So, how's the mix? In terms of the adoption of Watson, are there enough super techies out there who are absorbing this stuff or is it mostly packaged apps? Is it a mix? So it is a mix, but we know that data science skills are limited. I mean, they are coveted, right? And so those are the geeks, as you say, that are using the tool chain as a part of it. And we see that in a lot of customers and a lot of industries around the world. And then from a packaged app standpoint, the biggest use case of adoption is really around customer care, customer service, customer engagement, that kind of thing. And we see that as well, all around the world, all different industries, lots of great adoption. Watson Assistant is our flagship in that. So in terms of, if you think about these digital initiatives, everybody talks about digital transformation. Last few years, I mean, kind of started in 2016 in earnest. It's real when you talk to customers. And there was a ton of experimentation going on. It was almost like spaghetti, throw against the wall and see what sticks. Are you seeing people starting to place their bets on AI, narrowing their scope and really driving specific business value now? Or is it still kind of all over the place? Well, there's a lot of studies that say it's about 51% or so, still stuck in experimentation. But I would tell you in most of those cases even, they have a nice palette that's in production that's doing a part of the business. So, because people understand, while they may be interested in the sexiness of the technology, they really want to be able to get the business outcomes. So yes, I would tell you that things have kind of been guided, focused towards the use cases and patterns that are the most common. You know, and we see that, like I mentioned, customer care, we see it in how do you help knowledge workers? So you think of all those business documents and papers and everything that exists. How do you assist those knowledge workers, whether or not it's an attorney or an engineer or a mortgage loan advisor? So you see that kind of use case and then you see customers that are built in their own focused in on, you know, how do they optimize or automate or predict something in a particular line of business. So you mentioned Watson Assistant. So tell us more about Watson Assistant and how has that affected adoption? So Watson Assistant, as I said, is our flagship around customer care. And just to give you a little bit of a data point, Watson Assistant now through our public cloud, SAS version, converses with 82 million end users a month. So it's great adoption. And this is enabling customers, customers of our customers, to be able to get self-service help in what they're doing. And Watson Assistant, you know, a lot of people want to talk about it being a chat bot and you can do simple chat bots with it, but it's to sophisticated assistance as well. Cause it shows up to do work. It's there to do a task. It's to help you deal with your bank account or whatever ideas you're trying to do and whatever company you're interacting with. So chat bots is kind of a bit of a pejorative, but you're talking about digital assistant. I mean, it's like a super chat bots, right? And I saw a stat the other day that there's going to be, I don't know, 2025, whatever there's going to be more money spent on chat bot development or digital assistance than there is on mobile development. And I don't know if that's true or not, but it was kind of an interesting thing. So what are you seeing there? I mean, again, I think chat bots, people think of, oh, I got to talk into a bot, but a lot of times you don't know. So they're getting better. I liken it to fraud detection. 10 years ago, fraud detection was like six months later, you'll get a call. And so chat bots are just going to get better and better and better. And now there's this super category that maybe we can define here. What is that all about? That's right. And actually I would tell you, they can become the brain behind something that's happening. So just earlier today, I was with a customer and talking about their email CRM system and Watson Assistant is behind that. So chat bots aren't just about what you may see in a little window. They're really about understanding user intent, guiding the user through what they're trying to either find out or do and taking the action as a part of it. And that's why we talk about it being more than chat bots because it's more than a FAQ interchange. Yes, okay. So it's software that actually does, performs tasks. Yes. Probably could call other software to actually take action. That's right. We think of this as new systems of agency actually making sort of decisions. And then I guess the third piece of that is having some kind of human interaction and where we're appropriate, right? What are you seeing in terms of infusing humans into the equation? So, well, a couple of things. So one of the things that Watson Assistant will do is if it realizes that it's not the expert on whatever it is, then it will pass over to an expert and think of that expert as a human agent. And while it's doing that, so you may be in the queue because that human person is tied up, you can continue to do other things with it while you're waiting to actually talk to the person. So that's a way that the human is in the loop. I would tell you there's also examples of how the agents are being assisted in the background. So they have the interaction directly with the user, but Watson Assistant is helping them be able to get to more information quicker and narrow in on what the topic is. So you guys talk about the AI ladder, sort of Rob talked about that this morning. My first version of the AI ladder was building blocks. It was like data in AI, analytics, ML, and then AI on top of that. I said AI, data and IA, information architecture. Now you use verbs sort of to describe it, which is actually more powerful. Collect, organize, analyze, and infuse. Now infuse is like the holy grail, because that's operationalizing and being able to scale AI. What can you tell us about how successful companies are in infusing AI and what is IBM doing to help them? So I'm glad you picked up, first of all, that these are verbs and it's about action. And action leads to outcome, which is, I think, critical. And I would also tell you yes, infuse is the holy grail of the whole thing, because that's about injecting it into business processes, into workflows, into how things are done. So you can then see examples of how attorneys may be able to get through their legal prep process in just a few minutes versus 10, 15 hours on certain things. You can see conversion rates from a sales standpoint improve significantly. A number of different things. We've also got it as a part of supply chain optimization, understanding a little bit more about both inventory, but also where the goods are along the way. And particularly when you think of a very complicated thing, there could be a lot of different goods in various points of transit. You know, I was sort of joking, not joking, but mentioning Jeopardy at first, because a lot of people associate Watson with Jeopardy. I can't remember the first time I saw that. It had to be the mid part of the last decade. What was it? February of 2011. 2011, okay. I thought I even saw demos before that. I'm actually sure I did, back in some lab in IBM. And of course, the potential blew your mind. Right. I suspect you guys didn't even know what you had at the time. You were like, okay, we're going to go change the world. And you know, when you drive up and down one-on-one in Silicon Valley, you're like, oh, Watson, there's Watson that, you know, you got the consumer guys doing facial recognition, ad serving, you know, serving up fake news, you know, all kinds of applications. But IBM's trying to do something different. You're trying to really change business. Did you have any clue as to what you had at the time? And then how much of a challenge you were taking on and then bring us to where we are now and what do you see as the potential for the next 10 years? So, of course we had a clue. So let me start there. But with that, I think the possibilities of it weren't completely understood. There's no question in my mind about that. And what the early days were, were understanding, okay, what is that business application? What's the pattern that's going to come about as a part of it? And I think we made tremendous progress on that along the way. I would tell you now, you mentioned operationalizing stuff and you know, now it's about how do we help companies have it more throughout their company, through different lines of business? How does it tie to various things that are important to us? And so that brings in things like trust, explainability, the ethics of what it's doing, bias, detection and mitigation. And I actually believe a lot of that and the operationalizing it within the processes is where we're going to head going forward. Of course they'll continue to be advancements on the features and the capabilities, but it's going to be about that. I'm going to ask it, it depends questions. I know that's your answer, but at the macro, can machines make better diagnoses than doctors today? And if not, when will they be able to in your view? So I would actually tell you that today they cannot, but what they can do is help the doctor make a better diagnosis than she would have done by herself. And because it comes back to this point of, how the machine can process so much information and help the expert, in this case, the doctor's the expert, it could be an attorney, it could be an engineer, whatever, help that expert be able to augment the knowledge that he or she has as a part of it. So, and that's where I think it is, and I think that's where it will be for my lifetime. So there's no question in your mind that machines today, AI today, is helping make better diagnoses. It's just with an augmented or attended type of approach. And I want to talk about Watson anywhere. Okay, great. So we saw some discussion in the keynotes and some demos. My understanding is you can bring Watson anywhere to the data, don't have to move the data around. Why is that important? Give us the update on Watson anywhere. So first of all, this was the biggest requirement I had since I joined the Watson team three and a half years ago, was please can I have Watson on-prem? Can I have Watson in my company, data center, et cetera? And we needed to instead really focus in on what these patterns and use cases were, and we needed some help in the platform. And so thanks to Cloudpack for data and the underlying Red Hat OpenShift and Container Platform, we now are enabled to truly take Watson anywhere. So now you can have it on-premise, you can have it on the other public clouds. And this is important because like you said, it's important because of where your data is. But it's also important because the workloads of today and tomorrow are very complex. And what's on cloud today may be on-premise tomorrow, maybe in a different cloud. And as that moves around, you also want to protect the investment of what you're doing as you have Watson customized for what your business needs are. Do you think you timed it right? I mean, it kind of did. All those talk about multi-cloud now. You really didn't hear much about it four or five years ago. For a while, I thought you were trying to juice your cloud business. Say if you want Watson, you got to go to the IBM cloud. Was there some of that? Or was it really just, hey, now the timing's right where clients are demanding it in hybrid and multi-cloud and on-prem situation? Well, look, we know that cloud and AI go hand in hand. So there was a lot of positive with that. But it really was this technology point because had I taken it anywhere three and a half years ago, what would have happened is every deployment would have been a unique environment, a unique stack. We needed to get to a point that was a modern day infrastructure. If you will, and that's what we get now with a containers-based platform. So you're able to scale it such that every instance is in a snowflake that requires customization. That's right. So then I can invest in the enhancements to the actual capability that's there to do, not supporting multiple platform and stanchinations under the covers. Well, okay, so you guys are making that transparent to the customer. How much of an engineering challenge is that? Can you share that with us? You got to run on this cloud or that cloud or on-prem? Well, now, because of CloudPak for data and then what we have with OpenShift and Kubernetes and containers, it becomes, well, there's still some technical work my engineering team would tell you, it was a lot. But it's simple now. It's straightforward. It's a lot of portability and flexibility. In the past, it would have been every combination of whatever people were trying to do, and we would not have had the benefit of what that now gives you. And what's the technical enable there? Is it sort of open APIs, architecture that allows for the interconnectivity? So, but inside of Watson or the overall platform? The overall platform. So I would say at its core, it's what containers bring. Okay, really, so it's the marriage of your tech with the container wave. That's right, which is why the timing was critical now. So if you go back, yes, they existed, but it really hadn't matured to a point of broad adoption and that's where we are now. Yeah, the adoption of containers, Kubernetes, microservices, now it's on a very steep curve. All right, I'll give you last word on, big takeaways from this event. What are you hearing? What are some of the things you're most excited about? So first of all, that we have all of these clients and partners here and all the buzz that you see and that we've gotten. And then the other thing I would tell you is the great client examples and what they're bragging on because they are getting business outcomes and they're getting better outcomes than they thought they would achieve. Yeah, IBM knows how to throw an event. Beth, thanks so much for coming to the queue. Thank you, good to see you again. All right, great to see you. All right, keep it right there, everybody. We'll be back. This is theCUBE live from the IBM Data and AI Forum in Miami. We'll be right back.