 from our studios in the heart of Silicon Valley, Palo Alto, California. This is a CUBE Conversation. Hello everyone, welcome to this CUBE Conversation here in Palo Alto, California. I'm John Furrier, co-host of theCUBE. We are here with Inhi Chusa, general manager of IBM Watson customer engagement, former CUBE alumni. I think she's been on dozens of times. Great to see you again. Welcome to our Palo Alto studios. Yeah, great being here, John. So we haven't chatted in a while. IBM Think just happened. A little bit of a rainy event here in February. Interesting changeover since we last talked. But first, give an update on what you're up to these days. What group are you leading? What's new? Okay, well, first of all, I'm here based in California, which I'm excited about. And I lead our Watson West office, which is our Watson headquarters here on the West Coast in downtown San Francisco. And we hosted our Think conference. And at Think, I lead within IBM what we call our Watson customer engagement business unit, which is really the business applications of how we apply Watson and other disruptive tech to a line of business audiences, both SAS as well as on-premise software. And so really excited about the areas of applying AI and machine learning as well as blockchain to things like supply chain and logistics to order management to next generation of retail. A lot of new, exciting areas. You know, we've had many conferences over the years from big data as your career spanned across IBM. You had much more horizontal view of things now. You're horizontally scalable, as we say, in the cloud world. What's your observation of the trends these days? Because there's a lot of big way, obviously, that the wave that you guys announced was the IBM Watson Anywhere and the cloud private where Arvin and I had an amazing conversation that video went viral. This is now getting a big tailwind for IBM. What's your thoughts in general about the overall ecosystem? Because you're here in Silicon Valley, you're seeing the big waves, you got into the big data world, cloud is here, multi-cloud. What's your thoughts on the big mega trends? Yeah, that's a good question. I think the first chapter of cloud, everyone ran to public cloud. When you look at it through the lens of enterprise though, the hot topic right now in the second chapter is really about public cloud, not just public cloud, but multi-cloud, hybrid cloud, meaning whether it's private, public, it's about thinking about the applications and the nature of the applications. And regardless of where the data sits, what are the implications of actually getting work done through kind of new container services, new ways of microservices and the development of how APIs are integrated. And so the hot topic right now is definitely hybrid cloud and multi-cloud. And the work we've done to certify what we call IBM cloud private really enables us to not just take any business application to any cloud and our cloud as well, but actually to enable Watson and Watson-based applications also across multi-cloud environments. So chapter two, Ginny mentioned that in our keynote as well. I want to dig into that because we've been talking a lot about multi-cloud architecture. And one of the big debates has been in the industry, oh, don't pick a sole cloud. I've been writing a bunch of content about that, this DOD, Jet Ideal, and Amazon and Oracle fighting for it out there. But that's also happening in the enterprise. But the reality is everyone has multiple clouds. If you've got a Salesforce or if you got this and then the other thing, you probably have multiple clouds. So it's not so much sole cloud versus it is workloads having a cloud for the right job. And that seems to be validated at IBM Think in talking to the top technical people and in the industry, they all say pick the right cloud for the job. And we've heard that before in big data, pick the right tool for the job. So given that, workloads seem to be driving the demand for cloud. Since you're on the app side, how are you seeing that? Because the world's flipped. It used to be infrastructure and software enabled the app's capabilities. Now the workloads have infrastructure as code and with cloud, they're driving the requirements. This is a changeover. Talk about this. It is a big change and part of, I would say when people first ran to the cloud and a lot of the public cloud services were digital SaaS services where people were wanting to stitch multiple applications across clouds and that became a challenge. So in this next iteration that I'm seeing is really a couple of things. One is data gravity. So where does the data actually reside for the workload that's actually happening? Whether it's the transactions, whether it's customer information, whether it's product information. So that's one piece. The second piece is a lot more analytics, right? And the spectrum of analytics running from traditional warehouse capabilities to more let's say larger scale big data projects to full blown advanced algorithms and AI applications is people are saying, look, not only do I want to stitch these applications across multiple clouds, I also want to make sure that I can actually tap into the data to apply new types of analytics and derive new services and new values out of relationships, understanding of how products are consumed, and so forth. So for us, when we think about it is we want to be able to enable that fluid understanding of data across the clouds as well as protect and be thoughtful about the data privacy rights around it, compliance around GDPR, as well as how we think about the security aspects as well for the enterprise. That's a great point. And I think I want to drill down on the data piece coming to your background and data. Obviously it's going to be key in your job now. Obviously it's pretty obvious with Watson. But David Floyd, a Wikibon's research analyst just posted a taxonomy of hybrid cloud research report that laid out the different kinds of clouds you could have. There's edge clouds, there's all kinds of things from public to edge. So when you look at that, you're seeing, okay, the data plane is the critical nature of the cloud. Now, depending on which cloud architecture for the use case, the workload, whatever. The data plane seems to be this magical opportunity. AI is going to have a big part of that. Can you just talk about how you guys see that evolve it because obviously AI is killer part of your strategy. This data piece is interoperating across the clouds. Data management, governance, your smile is as a killer answer. Totally, yo, this is such a great setup. Actually, Ginny even said it in her keynote at Think, which was you can't have an AI strategy without a information architecture strategy, which is an IA strategy. And information architecture is all about what you said. It's data preparation, understanding the foundation of it, making sure you've got the right governance structure, the integration of it, and then actually how you apply the more advanced analytics on top. So information architecture and thinking about data aspects and all kinds of data. Majority of the data actually sits behind what I would say the traditional public firewall. So sits behind the firewalls of our enterprise clients, like 80 plus percent of it. And then many of the clients, we actually recently did a study with about 5,000 senior executives across many, many thousands of organizations. And 85% of them want to apply AI to improve their customer service, improve the way they engage their clients and their products and services. And so this is a huge opportunity right now for pretty much every organization to think through kind of their data strategy, their information architecture strategy as part of their overall AI strategy. So a question I got on Twitter comes up a lot and also on my notes here, I wanted to ask you is, how can companies increase transparency, trust and mitigate bias in AI? Because this comes up a lot and that's the question to come in from the community is, hey, I got my site, my apps running in Germany, I got users over there, I'm global, I have to manage compliance, I got all this governance now, I'm over my shoulders kind of pain in the butt, but also I don't want to have the software be skewed on bias and other things. And then I also get this whole Facebook dynamic going on where it's like, I don't trust people holding my data. This is a big, huge issue. You guys are in the middle of it. What's your thoughts? What's the update? What's the dynamic? What's the solution? So, this is a big topic. I think we could do a whole episode just on this topic alone. So trust and developing trust and transparency in AI should be a fundamental requirement across many, many different types of institutions. So first of all, the responsibility doesn't sit only with the technology vendors. It's a shared responsibility across government institutions, the consumers as well as the business leaders in terms of how they're thinking about it. The more important piece though, is when you think about the population that's available, that really understands AI and they're actually coding and developing on it, is that we have to think about the diverse population that's participating in the governance of it. Because you don't want just one tribe or one group that's coding and developing the algorithms or deciding the decision models. Like the nerds or the geeks. It's a social aspect, society aspect as well, right? Exactly. I actually just did a recent conversational series with Northwestern Kellogg's Business School around the importance of developing trust and transparency, not only in the algorithms themselves, but the methodology of how you think about culture and value and ethics come into play through different lengths, depending on the country you live in as you kind of reference, depending on your different values and religious backgrounds. It may be because of different institutional and or policy positions depending on the nature. And so there has to be a general awareness of this that's thoughtful. Now, why I'm so excited about the work we're doing at IBM is we've actually launched a couple new initiatives. One is what we call AI OpenSkill, which is really a platform and an opportunity to have the ability to begin to apply AI, see how AI operations and models function in production. We have a methodologies in terms of engaging, understanding fairness. So there's a 360 degree fairness kit, which is actually available in the open source world. There's a set of tools to understand and train people on recognizing bias. So even just definitions of what do you mean by bias? It could be things like group think, it could be you're just self selecting on certain data sets to reinforce your hypotheses. It could be unconscious levels. And it's not just traditionally socially oriented types of bias. It could be data bias, too. It could be data bias, right? Total machine generated biases in IoT world also. So contextual and behavioral bias is kind of kicking to play here. Yeah, but it starts with transparency, trust. It also starts with thoughtful governance. It starts with understanding in your position on policy around data privacy. And those things are things that should be educational conversations across the entire industry. How far along are we on the progress bar there? I mean, it seems like it's early, and we seem to be talking for a while, but it seems even more early than most people think. Still a lot more work to do. Your thoughts on where the progress bar is on this whole mashup of tech and social issues around bias and data. What's your thought? We're really at the early stages. And part of the reason we're at the early stages is I think people have so far really applied AI in very simple task oriented applications. The more what we call a broad AI, meaning multi-task workflow applications, they're starting. And we're also starting to see in the enterprise. Now, in the enterprise world, you can still have bias. So for example, when you talked about data bias, one of the simple examples I use is, think about like loan approvals. Well, if one of the criteria may be based on gender, well, you may have a sensitivity around the lack of women-owned business leaders, and that could be a scoring algorithm that says, hey, maybe it's a higher risk when in fact, it's not necessarily a higher risk. It's just that- It's just the sampling is off. It's the sampling is off, right? So that would be a detection to say, hey, maybe you have sensitivity around that data set because you actually have insufficient amount of data. So part of data detection and understanding bias is where you have a sampling of data that's incorrect, where your segmentation could be rethought, where it may just require an additional supervision or like decision-making criteria as part of your governance process. It sounds like a great area for young people to get involved, whether at their universities or curriculum, this kind of seems to be, whether it's political science and or data science kind of coming together. You kind of have a mashup. What's your advice to people watching that might be either in high school, college, or rethinking their career? Because this seems to be a hot area. It is a hot area. I would recommend it for every student at every age, quite frankly. And it's that we're at such an early stage that it's not too late to join and you're not too young nor are you too old to actually get in the industry. So that's point one. This is a great time for everyone to get involved. The second piece is I would just start with like online courses that are available, as well as participate in communities and companies. Like IBM, where we actually make available on number of our web based applications that you can actually do some online training and courses to understand the services that we have, to begin to understand the taxonomy and the language. So a very simple set would be like, learn the language of AI first. And then as you're learning coding, if you're more technically inclined, there's just a myriad of classes available. Final question before I move on to the topic around inclusion and diversity. Machine learning is impacting all verticals. I'm just interviewing, talking with Donna Dubinsky. She's got a company where it's Neuroscience and machine learning coming together. Machine learning is being impacted in all of our, we mentioned just basic data bias and those things, machine learning can help there. Machine learning meets blank, every vertical, every market is being impacted machine learning, which will trigger some of the things you're seeing on the app side. Your thoughts looking at where you've come from in your career at IBM to now, just the evolution of what machine learning has enabled. Your thoughts on the impact of machine learning. Oh, it's exciting. And I'll give you a real simple example. So one of the great things my own team actually did was apply machine learning to, everyone loves the holiday shopping period, right? Between Thanksgiving to New Year's. So we actually developed what we call Watson order optimizer. And one of my favorite brands is REI, so the recreational equipment incorporated company. They actually applied our Watson order optimizer to optimize in real time, the best place that, let's say you want to order a kayak or a t-shirt or a hiking boot, but the best way to create the algorithms to ship from different stores and shipping from stores for most retailers is a high cost variable because you don't know what the inventory positions are, you don't necessarily know the movement of traffic into that store. You may not even know what the price promotions are. So what was exciting about putting machine learning algorithms to this was we could actually curate things like shipping and tax information, inventory positions of products in stores, pricing, a movement of goods as part of that calculation. So this is like a set of business rules that are automatically developed using Watson in a way that would be almost impossible for any human to actually come up with all of the possible business rules, right? Because it's such a complex situation. And then you're trying to do it at the peak time, which is like Black Friday, Cyber Monday weekend. So we were able to actually apply Watson machine learning to create the business rules for when it should be shipped from a warehouse or a particular store in order to meet the customer requirement, which is the fulfillment of that brand experience or the product experience. So my view is there are so many different places across the industry that we could actually apply machine learning to. And my team is really excited about what we've been doing, especially in the next generation of supply chain. And it's also causing students to be really attracted to computer science, both men and women. My daughter, who's a senior at Berkeley, is interested in it. So you're starting to see that the impact of machine learning is hitting all mainstream, which gives a good segue to my next question. You know, we've been very passionate. I know it's one of your passions is inclusion and diversity or diversity and inclusion. Yeah, there's always a debate. Is it D before I or I before D? Some say inclusion and diversity or diversity and inclusion, it's all the same thing. There's just a lot of effort going on to bring the tech industry up to par with the reality of the world. And so you guys have a study out, I got a copy here. Talk about this study, women in leadership and the priority paradox. Talk about the study, what was behind it and what are some of the findings? Sure, and I'm excited that your daughter, who's a senior in college, is going to be another woman that's entering the workforce and especially being in tech. So the priority paradox is that we actually looked at over 2,300 organizations. These are some of the top institutions around the world that are curating and attracting the best talent and skills. Now, when you look at that population, we were surprised to find out that you would think by 2019, right? Through 2018, that only 18% of those organizations actually had women in senior leadership positions. And what I categorize as senior leadership positions are in the C-suite as vice presidents, maybe senior executives or senior managers, director level folks. So that's one piece, which is, wow, given the size and the state at where we are in the industry, only 18%, we could do better. Now, why do we believe that? The second piece is that you want the full population of the human capacity to think and creatively solve some of the world's biggest complex problems. You don't want a small population of the world trying to do this. So the second piece of the paradox, which was the most surprising is that 79% of these companies actually said that formalizing or prioritizing gender, fostering that kind of inclusive culture was not a business priority and that they had a harder time actually mapping that gap. Now, in the study, what we actually discovered though was those companies that did make it a priority actually had first mover advantage. And making a priority is quite simple. It's about understanding how to create that inclusive culture to allow different perspectives and different experiences to be allowed in the, you know, co-creation and development. So first mover advantage in terms of the statistics or the business performance? Actual business performance. So even though 80% of the organizations that we interviewed actually said that they've not made it a business priority, the 20% that did, we actually saw higher performance in their outcomes in terms of their business performance. So there's actually a business benefit to it. That your point is the first mover advantage is saying those companies that actually brought in the leadership to create that different perspective had higher performance. Absolutely. Well, I know, we've talked about this before. One of the things I always say is that, you know, tech is now mainstream and it's 18% of the target audience of tech isn't the market. It's 50, 50 or 51, say 51% women men. So who's building the products for half the audience? So, so again, this doesn't make any sense. So this is a good, good statistic. It is. And if you think about the students that are actually graduating out of graduate school recently, there's actually more women graduating out of grad school than men. And when you think about that population, that's now entering the workforce. And what's actually happening through the pipeline, I think there's got to be thoughtful focus and programmatic improvements across the industry around how to develop talent and make sure that different companies and organizations can move, like you said, problem solved for creating new products that actually serve the world, not just serve certain populations, but also do it in a way that's thoughtful about kind of the makeup. And the mainstream of tech obviously makes it more attractive. I mean, you're seeing a lot more women thinking about machine learning like my daughter and others. The question is, how do they come in and not lose their footing, mentorship? So what are the priorities that you see the industry needs to do? What are some of the imperatives to keep the pipeline and keep all the mentoring? I see mentoring is hot. We see the networks being built. Yeah, mentoring is huge. What's your thoughts on the best practices that you've been involved in? Some of the best practices, we've actually done a number within IBM. We've done a program called Tech Reentry. So women that have like decided to come back into the tech workforce, we actually have a 12-week internship program to do that. Another is a big initiative that we have around P-Tech, which is the next generation of workers aren't just going to have a formal college and or PhD master's type degrees. The next generation, what we're calling is not necessarily a white collar or blue collar, what we're calling it is new collar, meaning these are students that are able to combine their equivalent of a high school degree and early college education in one to be kind of, if you think about it, next generation of technical vocational schools, right, that quickly enter the workforce are able to do jobs in terms of web development, in terms of cloud management, cloud services. It could be a next generation of- You have skill gap opportunity. There's a big opportunity for people. It is, and we're seeing great adoption. We've seen it on a number of states across the US. This is an effort that we partner with the states and the governors of each state because public education, it's got to be done in a systematic way that you can actually sustain it for many, many years. And this is something that we were excited about championing in the state of New York first. The reaction program, and other things I've always told my friends, look at the technology is so new now, you could level up a lot faster than there's not that linear school kind of mentality. You don't need eight years to learn something. You could literally learn something pretty quickly these days because the level, and the gap between you and someone else is so short now, because it's all new skills. It's true, and the, you know, we talk about digital disruption through the lens of businesses, but there's a huge digital disruption through the lens of what you're talking about, which is our individual kind of development and talent, and the ability to learn through so many different channels that's available now, and the focus around micro degrees, micro skills, micro certifications. I mean, there's so many ways for everyone to get involved, but I really do encourage everyone across every industry to have some knowledge and basis and understanding of tech, because tech will redefine how services and products are delivered across every category. And that's not male or female, that's just everyone. Again, back to technology for good. We can solve technology problems. You guys have been doing it at IBM, solve technology problems, but now the people problem is about getting people empowered, all gender, races, et cetera. The people gap, getting the skills, getting them employed, working for clouds. This is an opportunity. It's a huge opportunity. I think this is an exciting time. We feel like we're entering this next phase of what I call chapter two of cloud, is a chapter two of digital reinvention of the enterprise, digital reinvention of the individual, actually. And it's an opportunity for every country, every population group to get involved in so many new and creative ways. And we're at the early foundation stages in terms of both AI development, as well as new capabilities like blockchain. So it's an exciting time for everyone. Well, as a whole nother topic, we'll have to bring you back in here. Great to see you. In fact, welcome to Palo Alto, our first time in our studio. I'd like to get you, let's co-host something together. Me and you. We're two of series, John and Inhi. I would love that, that would be fun. I'm excited to be here, so thank you. You can drop by our studio anytime now that you live in Palo Alto, we're neighbors. Inhi, you saw here, general manager, IBM Watson, customer agent, friend of the cube, here inside our studios in Palo Alto. I'm John Furrier, thanks for watching.