>> Live from Boston, it's the CUBE. Covering IBM Chief Data Officer Summit. Brought to you by IBM. >> Welcome to the CUBE's live coverage of IBM Chief Data Officer Summit here in Boston, Massachusetts. I'm your host, Rebecca Knight along with my co host, Paul Gillin. We're starting our coverage today. This is the very first day of the summit. We have two guests, Caitlin Halferty, she is the AI accelerator lead at IBM, and Sonia Mezzetta, the data governance technical product leader. Thank you both so much for coming on the CUBE >> Thanks for having us. >> So this is the ninth summit. Which really seems hard to belief. But we're talking about the growth of the event and just the kinds of people who come here. Just set the scene for our viewers a little bit, Caitlin. >> Sure, so when we started this event back in 2014, we really were focused on building the role of the chief data officer, and at that time, we know that there were just a handful across industries. Few in finance banking, few in health care, few in retail, that was about it. And now, you know, Gartner and Forrester, some industry analysts say there are thousands across industries. So it's not so much about demonstrating the value or the importance, now, it's about how are our Chief Data Officers going to have the most impact. The most business impact. And we're finding that they're really the decision-makers responsible for investment decisions, bringing cognition, AI to their organizations. And the role has grown and evolved. When we started the first event, we had about 20, 30 attendees. And now, we get 140, that join us in the Spring in San Francisco and 140 here today in Boston. So we've really been excited to see the growth of the community over the last four years now. >> How does that affect the relationship, IBM's relationship with the customer? Traditionally, your constituent has been the CIO perhaps the COO, but you've got this new C level executive. Now, what role do they play in the buying decision? >> There was really a lot of, I think back to, I co-authored a paper with some colleagues in 2014 on the rise of Chief Data Officer. And at that time, we interviewed 22 individuals and it was qualitative because there just weren't many to interview, I couldn't do a quantitative study. You know, I didn't have sample size. And so, it's been really exciting to see that grow and then it's not just the numbers grow, it's the impact they're having. So to you questions of what role are they playing, we are seeing that more and more their scope is increasing, their armed and equipped with teams that lead data science, machine learning, deep learning capabilities so they're differentiated from a technology perspective. And then they're really armed with the investment and budget decisions. How should we invest in technology. Use data as a strategic corporate asset to drive our progress forward in transformation. And so we've really seen a significant scope increase in terms of roles and responsibilities. And I will say though, there's still that blocking and tackling around data strategy, what makes a compelling data strategy. Is is the latest, greatest? Is it going to have an impact? So we're still working through those key items as well. >> So speaking of what makes this compelling strategy, I want to bring you into the conversation Sonia, because I now you're on the automated metadata generation initiative, which is a big push for IBM. Can you talk a little bit about what you're doing at IBM? >> Sure. So I am in charge of the data governance products internally within the company and specifically, we are talking today about the automated metadata generation tool. What we've tried to do with that particular product is to try to basically leverage automation and artificial intelligence to address metadata issues or challenges that we're facing as part of any traditional process that takes place today and trying to do curation for metadata. So specifically, what I would like to also point out is the fact that the metadata curation process in the traditional sense is something that's extremely time-consuming, very manual and actually tedious. So, one of the things that we wanted to do is to address those challenges with this solution. And to really focus in and hone in on leveraging the power of AI. And so one of the things that we did there was to basically take our traditional process, understand what were the major challenges and then focusing on how AI can address those challenges. And today at 4 p.m. I'll be giving a demo on that, so hopefully, everybody can understand the power of leveraging that. >> This may sound like a simple question, but I imagine for a lot of people outside of the CIO of the IT organization, their eyes glaze over when they hear terms like data governance. But it's really important. >> It is. >> So can you describe why it's important? >> Absolutely. >> And why metadata is important too. >> Absolutely. Well, I mean, metadata in itself is extremely critical for any data monetization position strategy, right. The other importance is in order to derive critical business insights that can lead to monetary value within a company. And the other aspect to that is data quality which Interpol talked about, right? So, in order for you to have the right data governance, you need to have right metadata in order for you to have high level of data quality can, if you don't and you're spending a lot of time cleaning dirty data and dealing with inefficiencies or perhaps making wrong business decisions based on bad data quality, it's all connected back to having the right level of data governance. >> So, I mean, I'm going to also go back to something you were talking about earlier and that's just the sheer number of CDOs that we have. We have statistic here, 90% of large global companies will have the CDO by 2019. That's really astonishing. Can you talk a little bit about what you see as sort of the top threats and opportunities that CDOs as grappling with right now. >> And let me make this tangible. I'll just describe my last two weeks, for example. I was with the CDO in person in Denver of a beer company, organization, and they were looking at some MNA opportunities and figuring out what their strategy was. I was at a bank in Chicago with the head of enterprise data government there, looking at it from a regular (mumbles) perspective. And then I was with a large multinational retail organization with their CDO and team figuring out how did they work at a sort of global scale and what did they centralize at enterprise data level. And what did they let markets and teams customize out in the field, out in the GOs. And so, that's just an example of, regardless of industry, regardless of these challenges, I'm seeing these individuals are increasingly responsible for those strategic decisions. And oftentimes, we start with the data strategy and have a good discussion about what is that organization's monetization strategy. What's the corporate business case? How are they going to make money in the future and how can we architect the data strategy that will accelerate their progress there? And again, regardless of product we're selling or retail, excuse me, our industry, those are the same types of challenges and opportunities we're grappling with. >> In the early days there was a lot of questions about the definition of the role and those CDOs set in different departments and reported to different people, are you seeing some commonality emerge now about how this role, where it sits in the organization, and what its responsibilities are? >> It's a great question, I get that all the time. And especially for organizations that recognize the need for enterprise data management. They want to invest in a senior level decision-maker. And then it's a question of where should they sit organizationally? For us internally, within IBM, we report to our Chief Financial Officer. And so, we find that to be quite a compelling fit in terms of budget. And visibility into some of those spend decisions. And we're on par in peers with our CIO, so I see that quite a bit where a Chief Data Officer is now on par and appear to the CIO. We tend to find that when it's potentially buried in the CIO's organization, you lose a little of that autonomy in terms of decision-making, so if you're able to position as partners and drive that transformation for your organization forward together, that can often work quite well. >> So that partnership, is it, I mean ideally, it is collaborative and collegial, but is it ever, are there ever tensions there and how do you recommend the companies get over, overcome those obstacles? >> Absolutely, in the fight for resources that we all have, especially talent and retaining some of our top talent, should that individual or those teams sit within a CIO's organization or a CDO's organization? How do we figure that out? I think there's always going to be the challenge of who owns what. We joke, sometimes, it feels like you own everything when you're in the data space, because you own all of the data that flows through, all your business processes, both CDO-owned and corporate HR's supply chain finance. Sometimes it feels you don't own anything. And so we joke that it's, you have to really carve that out. I think the important part is to really articulate what the data strategy is, what the CDO or enterprise data management office owns from a data perspective and then building up that platform and do it in partnership with your CIO team. And then you really start to be able to build and deploy those AI applications off that platform. That's what we've been able to see, so. >> I want to go back to something Sonia said this morning during the keynote, you talked about IBM's master metadata list catalog unifying your organization around a certain set of terms. There's 6,000 terms in that catalog. Now, how did you arrive at 6,000? And what are some rules for an organization trying to do something like that? How defined, how small should that sub-terms be? >> Sure. Well, we started off with a traditional approach which is probably something that most companies are familiar with these days. The traditional process was really just based on basically reaching out to a large number of subject matter experts across the enterprise that represent in many different data domains such as customer, offering, financial, etc. And essentially having them label this data, specifically with the business metadata that's used internally across a company. Now, another example to that is that there are different organizations across the company. We are a worldwide company. And so, what one business might call a particular piece of data, which is customer, another might call it client. Which really ended up being this very large list of 6,000 business terms which is what we're using internally. But one thing that we're trying to do to be able to kind to basically connect the different business terms is leverage knowledge management and specifically ontological relationships to be able to link the data together and make it more reasonable and provide better quality with that. >> What are the things that you were talking about, Interpol was talking about on the main stage too during the keynote, was making sure that the data is telling a story because getting by in is one of the biggest challenges. How do you recommend companies think about this and approach this very big daunting task? >> I'll start and then I'm sure you have a perspective as well. One of the things that we've seen internally and I work with my client on, is every project we initiate, we really want strong sponsorship from the business in terms of funding, making sure that the right decision-makers are involved. We've identified some projects for example, that we've been able to deploy around supply chains. So identifying the risk on our supply chain processes. Some of the risks in sites, we're going to demo a little bit later today. The AMG work that Sonia's leading. And all of those efforts are underway in partnership with the business. One of my favorite ones is around enabling our sellers to better understand information about, and data, about the customers. So like most organizations, customer data is housed in silo systems that don't necessarily talk well with each other, and so it's an effort to really pull that data together in partnership with our digital sellers and enable them to then pull up user interface, user-friendly, an app where they can identify and drill down to the types of information they need about their customers. And so our thought and recommendation based on our experience and then what I'm seeing is really having that strong partnership with the business. And the contribution funding, stakeholder involvement, engagement, and then you start to prioritize where you'll have the most impact. >> You did a program called the AI accelerator. What is that? >> We did, so when we stood up our first chief data office, it was three years ago now, we wanted to be quite transparent about the journey of driving cognition through our enterprise. And we were really targeting those CDO and processes around client master product data and then all of our enterprise processes. So that first six months was about writing the data strategy and implementing that, next we spent a year on all of our processes, really mapping out, we call it journey mapping, I think a lot of folks do that, by process. So HR, supply chain, identifying ways. How it's done today, how it will be done in a cognitive AI like future state. And then also, as we're driving out those efficiencies in automation, those reinvestment opportunities to free up that money for future initiatives. And so that was the first year, year and a half. And now, we're at the point where we've evolved far enough along that we think we're learned some lessons on the way and there's been some hurdles and stumbling blocks and obstacles. And so a year ago, we really start a cognitive enterprise blueprint and that was really intended to reflect all of our experiences, driving that transformation. A lot of customer engagements, lot of industry analysts feedback as well. And now we formalized that initiative. So now I have a really fantastic team of folks working with me. Subject matter domain expertise, really deep in different processes, solutions, folks, architects. And what we can do is pull together the right breadth and depth of IBM resources. Deploy it, customize it to customer need and really, hopefully, accelerate and apply a lot of what we've learned, lot of what the clients have learned, to accelerate their own AI transformation journey. >> But AI, IBM is the guinea pig and it showcase. And so you're learning as you go and helping customers do that too. >> Exactly and we've now built our platform, deployed that, as we mentioned, we've got about 30,000 active users, active users, using our platform. Plan to grow to 100,000. We're seeing about 600 million in business benefit internally from the work we've done. And so we want to really share that and do some good, best practice sharing and accelerate some of that process. >> IBM used the term cognitive rather than AI. What is the difference or is there one? >> I think we're starting actually to shift from cognitive to AI because of that exact perspective. AI, I think is better understood in the industry, in the market and that's what's resonating more so with clients and I think it's more reflective of what we're doing. And our particular approach is human in the loop. So we've always said rather than the black box sort of AI algorithms running behind the scenes, we want to make sure that we do that with trust and transparency, so there's a real transparency aspect to what we're doing. And the other thing I would notice, we talk about sort of your data is your data. Insights derive from that data is your insights. So we've worked quite closely with our legal teams to really articulate how your data is used. If you engage and partner with us to drive AI in your enterprise, making sure we have that trust and transparency (mumbles) clearly articulated is another important aspect for us. >> Getting right back to data governance. >> Right, right, exactly. Which is our we've come full circle. >> Well Caitlin and Sonia, thank you so much for coming on the CUBE, it was great. Great to kick off this summit together. >> Great to see you again, as always. >> I'm Rebecca Knight for Paul Gillin, stay tuned for more of the CUBE's live coverage of IBM CDO Summit here in Boston. (techno music)