 Live from Boston, it's theCUBE, covering IBM Chief Data Officer Summit. Brought to you by IBM. Welcome back everyone to theCUBE's live coverage of the IBM CDO Summit here in beautiful Boston, Massachusetts. I'm your host, Rebecca Knight, along with my co-host, Paul Gillan. We have two guests for this segment. We have Susan Wagner, who is the VP Data, Artificial Intelligence and Governance at Deutsche Telecom, and Madhu Kochar, who's the Vice President Analytics Product Development at IBM. Thank you so much for coming on the show. Thank you. Happy to be here. Susan, you're coming to us from Berlin. Tell us a little bit about what, it's a relatively new job title, and Paul was marveling before the cameras were rolling, that you have artificial intelligence in your job title. Tell us a little bit about what you do at Deutsche Telecom. So we have a long history working with data, and this is a central role in the headquarter, guiding the different data and artificial intelligence activities within Deutsche Telecom. So we have different countries, different business units. We have activities there. We have already use case catalog of 300,000 cases there. And from a central point, we are looking at it and saying, how are we able really to get the business benefit out of it? So we are looking at the different product, the different cases, and looking for some help for the business units, how to scale things. For example, we have a case, we implemented in one of our countries, it was about the call center to predict if someone calls the call center, if this is a problem we would never have on Deutsche Telecom, but it could happen, and then we open a ticket, and we are working on it, and then we're closing that ticket, but the problem is not solved, so the ticket comes again, and the customer will call again. And this is very bad for us, bad for the customer, and we did an AI project there, predicting what kind of tickets will come back in future. And this we implemented in a way that we are able to use it not only in one country, but really give it to the next country, so our other business units, other countries, can take the code and use it in another country. That's one example. Wow, wow. How would you define artificial intelligence? There's someone who has it in your job time. That's sometimes a very difficult question, I must admit. I'm normally, if I would say from a scientific point, it's really to have a machine that works in fields and did everything like a human. If you look now at the hype, it's more about how we learn, how we do things, and not about, I would say, it's about robotic and stuff like that, but it's more how we are learning. And the major benefit we are getting now out of artificial intelligence is really that we are able now to really work on data. We have great algorithm, a lot of progress there, and we have the chips that develop so far that we are able to do that. It's far away from things like a little kid can do, because a little kid can just, you show them an apple and then it knows an apple is green. It, it's very. But a little kid can't open a support ticket. Yeah, but that's very special. So in very special areas, we are already very, very good in things, but this is an area, for example, if you have an algorithm who is able, like we did, to predict this kind of tickets, this algorithm is not able at the moment to say this is an apple and this is an orange. So you need another one. So we are far away from really having something like a general intelligent there. Madhu, I want to bring you into this conversation a little bit, just in terms of what Susan was saying, the sort of the shiny newness of it all, where do you think we are in terms of, in terms of thinking about the data, getting in the weeds of the data, and then also sort of the innovations that we all dream about, really impacting the bottom line and making the customer experience better and also the employee experience better. Yeah, so from IBM perspective, and especially coming from data and analytics, very simple message, right? We have what we say, your letter to AI. Everybody, like Susan and every other company who is part of doing any digital transformation or modernization is talking about AI. So our message is very simple. In order to get to the letter of AI, the most critical part is that you have access to data. You can trust your data. So this way you can start using it in terms of building models, not just predictive models but prescriptive and diagnostics, everything needs to kind of come together, right? So that is what we are doing in data and analytics. Our message is very, very simple. The innovations are coming in from the perspectives of machine learning, deep learning and making. And to me, that all equates to automation, right? A lot of this stuff, data curation, I think you can, Susan, how long and how manual the data curation aspects can be. Now with machine learning, getting to your letter of AI, you can do this in a matter of hours, right? And you can get to your business users. You can, if your charn model, if your clients are not happy, your fraud, you have to detect in your bank or retail industry, it just applies to all the industry. So there is tons of innovation happening. We just actually announced a product earlier called IBM Cloud Private for data. This is the analytics platform which is ready with data, built in governance, to handle all your data curation and be building models, which you can test it out, have all the DevOps, then push it into production. Really, really trying to get clients like Deutcher Kellogg to get their journey there faster. Very simple message. We've heard from many of our guests today about the importance of governance, of having good quality data before you can start building anything with it. What was that process like? How is the, what is the quality of data like at Deutsche Telekom and what work did it take to get it in that connection? Yeah, it's, so data quality is a major issue everywhere because as Maju, this is one of the essential things to really get into learning. If you want to learn, you need the data. And we have in the different countries different kind of majorities. And what we are doing at the moment is that we are really doing it case by case because we cannot do everything from the beginning. So you start with one of the cases, looking what to do there, how to define the quality. And then if the business asks for the next case, then you can integrate that. So you have the business impact, you have demand from the business, and then you can integrate the data quality there. And we are doing it really step by step because to bring it to the business from the beginning, it's very, very difficult. You had mentioned one of the new products that you announced just today. What are some of the most? We announced it in May. Oh, okay, yeah, I'm sorry. Still new. In terms of the other innovations in the pipeline, I mean, this is such a marvelous and exciting time for technology. What are some of the most exciting developments that you see? I think the most exciting, especially if I talk about what I do day in, day out, everything revolves around metadata, right? It used to be not a very sticky term, but it is becoming quite sexy all over again, right? And all the work in automatic metadata generation, understanding the lineage, where the data is coming from, how easy we can make it to the business users. Then all the machine learning algorithms which we are doing in terms of our prescriptive models and predictive maintenance is such a huge thing. So there's a lot of work going on there. And then also one of the aspects is how do you build once and run anywhere, right? If you really look at the business data, it's behind the firewalls, it's in multi-cloud. How do you bring solutions which are going to be bringing all the data doesn't matter where it resides, right? And so there's a lot of innovations like that which we are working and bringing in on to our platform to make it really simple story, make data easy access which you can trust. One of the remarkable things about machine learning is that the leading libraries have all been open sourced, Google, Facebook, eBay, others have open sourced, they're libraries. What impact do you think that has had on the speed with which machine learning has developed? Oh, just amazing, right? I think that gives us that agility to quickly able to use it, enhance it, give it back to the community. That has been one of the tenants for I think how everybody is out there moving really, really fast. Open source is going to play a very critical role for IBM and we are seeing that with many of our clients as well. What tools are you using? Are we using different kind of tools depending on the department. So the data scientists like to use our patents and Scala, they always use it. But we are using a lot like the Jupiter notebook, for example, to have different kind of code in there. We have in one of our countries the classical things like that's there and the data scientists working with. With that one, all we have the Cloudera workbench to really bringing things into the business we have and some business and IBM things integrated. So it really depends a little bit on the different. And that's a little bit the challenge because you really have to see how people working together and how do we really get the data, the models, the sharing right. And then also the other challenges that all the CDOs face that we've been talking about today, the getting by in, the facing unrealistic expectations of what data can actually do. I mean, how would you describe how you are able to work with the business side as a chief in the, working in the chief data office? So what I really like and what I'm always doing with the business that we are going to the business and doing really a joint approach, having a workshop together like the design thinking workshop with the business and the demand has to come from the business. And then you have really the data scientists in there, the data engineers, best to have the operational people in there and even the controlling, not all the time, but that it's really clear that all people are involved from the beginning and then you are really able to bring it into production. That's the term of data ops, right? That's starting to become a big thing. DevOps was all about agility. Now data ops bring all these various groups together and yeah, I mean, that's how you really move forward. So for organizations, that's both of you. For organizations that are just beginning to go down the machine learning path that are excited by everything they've been hearing here, what advice would you have for them? They're just getting started. I think if you're just getting started to me, the long pull item is all about understanding where your data is, right? The declaration, I have seen over and over again, everybody's enthusiastic, they love the technology, but it just doesn't progress fast enough because of that. So invest in tooling where they have automation with machine learning, where they can quickly understand it, right? Data virtualization, nobody's going to move data, right? They're sitting in bedrock systems. Access to that, which I call dark data, is important because that is sometimes, you know, your golden nugget because that's going to help you make the decisions. So to me, that's where I would focus first. Everything else around, it just becomes a lot easier. So, do you have a best practice too? Focus on really bringing quick impact on some of the cases because the management needs success, so you need some kind of quick success and then really working on the basics. Like Manjul said, you need to have access of the data because if you don't start to work on that, it will take you every time, like half a year, we have some cases where it took to the finance department to have a year to really get all that kind of data and you have to shorten that for the future, but you need the first quick wins, you need to do both. Excellent advice. Great. Well, Susan and Madhu, thank you so much for coming on theCUBE. It's been great having you. Yeah, thank you. Thank you. Thank you. I'm Rebecca Knight for Paul Gillan. We will have more from the CUBE's live coverage of the IBM CDO just after this.