My. Name is lloyd bird I'm the vice president of application. Responsibilities, are really the whole application, set for the company, those that are running in contact. Centers those are running in our corporate. Corporate. Support systems, as well along, with technical, solution. Immerse. Maybe. Just to share a couple of things about Convergys. To. Put. It into context. We're. Greater than, 115,000. Employees. We're. A worldwide, company, we're, in customer, care outsourcing, where the number to customer, care outsourcing, company in the world, we. Service greater, than 50 percent of, the fortune 500 in, our, customer, base. 58. Languages. Nearly. Three billion in revenue, and. A. Lot of contacts, 8 billion contacts, a year, so. You already start to see big numbers coming. Into play especially around employees. And around customer interactions, that, plays into our, journey to. GDP. Are, in. Amia specifically. So, when we say GDP, our were really talking about global. But. What. Really, really got, us going in, terms of developing, a strategy was EU, GDP. Our and having. To solve for the EU, directive. That. Came, out and, there, were 43, locations. 15, countries, 32. Languages, close. To 20,000, employees across. Amia. Not. All of those in the EU some. Of those in Tunisia. And, Cairo, and in. Nearshore. Amia. Footprints. But all servicing, EU customers, any EU customer, base and with. That we had a large amount of. Considerations. To figure out what to do. So. About I guess. May, of, 2017. About. A year before the directive, was going to go in place we, really started. Kind of hunkering down to figure out what we were going to do and how we were going to do it in. Order to satisfy. The directives, that we're coming, so. With that we kind. Of started, from a consultancy, perspective. And said what are all the things that we need to look at when. We're thinking about gdpr, and, in specifically. In this case EU, GDP, our and. We came across eight, we. Had then we knew we had to get with, our compliance, team we had to get policies. And procedures, and all of those things kind of lined up but. We, also knew, that, we. Were. Going to have to get a number of IT. Procedures. And policies, directly. Aligned. Including. Some. Governance, processes. That we were working with clients. All. Of our systems have access, controls, they all had backup, policies. Backup. Retention. The. Definitions. Etc, but, some of those didn't match what were needed for ugdp are and. We. Had to get just. Make sure all of our processes and change and so forth were lined up the. Tough ones came around encryption. Around. Data at rest, around. Data masking, and then. Probably the toughest, came. Around with, the, which, really is the key to it all how are you actually going to do data subject writers requests across. All your systems so we, fundamentally came, out of this period, of May. To. September, ish, of. 2017. With a realization. That we. Didn't have any good ways to handle data subject, requests we're. Still a little fuzzy about what we needed to do and. We. Didn't. Have a real, good way of discovering, all of that information as, well so we probably, started getting a little nervous about that time. So. That got us to September. Of 2017. We're. Sitting. There looking at the clock in, the counter we're like eight months to go live we've. Been thinking about some things and some ideas including. Graph technology. Of using. That because the, problems, was starting to get you know very very multi-dimensional, a. Lot of of. Data. And in various sources and so for so we started scoping, and, we. Started, by looking across our application, set we had 120, apps, that we really needed to look at some of those were things we built some were hybrid, some were third-party, apps and, we. Needed to look at those. We. Had a whole bunch of internal. Collaboration, and. Storage. Things. To look at that included, email systems, including, email back to clients. Included internal, emails SharePoint. You. Know on we, have a SharePoint, on-prem we have shared drives just to give you a sense of context with 8 terabyte, of SharePoint. To. Look at in this particular, context. So. That added, more pieces. To, it. Certainly. We had data centers, to think about we had backup, storage to think about infrastructure. Pieces that, we had to look at and. Then there were all of these operational. Parameters that. Started, to emerge. 106. Customers, across. That amia footprint. When. You really look at pieces of customer, business not, just customers, we'll call those lines of business that number gets up Prosser to 4 to 500, unique. Pieces of business large. Pieces of business with customers across 32, languages, 15, countries, 43. Sites and 19,000. Employees. So. Throw. In some contracts, throw in some regulations. All. Those bubbles, start, to look like things. That could be connected, with, graph. Technology. And so. We. More and more were thinking, that, some, other method, that we had used traditionally, was going to make sense in trying to figure out all the interconnectivity.
Of This. Data and and. That was kind of our I. Think, aha moment, was at that point. Around, September, of last year that. We said we got to figure out a new way of thinking, about this, it took us four months to analyze, this, data just, to get up a good. Handle, on what some of these relationships were so. We had originally thought okay, we're gonna build okay. Because. We like to build things so, we were gonna build and we. Have a development, organization. Even. If we're gonna use new technology, around graph. Technology. We were going to build it ourselves. We kind of looked out on the market, we couldn't really find anything, we, had some big box companies, that were coming to us and say will consult with you and charge. You a whole bunch of money and then we can together figure out what needs to be done but. Not really a solution, so. That didn't feel right and it, it really did seem like at this point everybody, was trying to figure out this problem on their own and in parallel, to each other and. Everybody. Was kind of designing, as we were going, so. We started, looking at, okay what even. Building graph models, and I'll show you some in here in a minute of. What the data would look like and so forth and, we. Started looking at the tough stuff well I'll call the tough stuff and, that. Especially. Was finding, data that was non structured, so. We we, got pretty good handle on our applications. We've got a pretty good handle on our data. Warehousing, on, our kind. Of key systems. And where the data are we use what's called a global operating, model so we didn't have a lot of variation. From region to region we, have good amount of consistency but. The, unstructured. Data was, was, throwing, us for a bit of a loop in terms of how do we find. Customers. That, might be in that unstructured, data or, employees, since the regulation, hit employees, as well then. It was what are we going to do with these data subject, right to actions there's eight listed, there of, different. Things you have to do it's not just you. Know a race data, you, know you've got to be able to at least report, on it you've got to be able to.
At. Least disclose. It in some cases it, goes all the way to erasure. But but other kind, of pieces in between. Reporting. And compliance was, going to be a bit of a problem for us because we have to build some, UI. To, capture these. Requests. For data subject rights we have to report back in the, manner that the EU said you have to port back on, these, particular, things we didn't have any of that we were going to build that with. Something was kind of looming out there and then there was all these apps and I'll go into some architecture. Here a minute our enterprise architecture, but. We. Were going to change all those apps and there's one problem with all of these apps. These. Apps were all built to put, data in records. At a time. None. Of these apps were built to take data out records, at a time and. Or. Manipulate. Individual. Records, within these apps and, so, even though we had a good handle on them technically. We didn't have a good way to. Address this, problem. So, here, we got the issue of finding it and, we got the issue - what to do on the, back end to try to deal with so. Our beginning state was we had a whole bunch of these these systems. There, they. Actually, are connected, very well especially. To, data, warehouse, structures, to, kind. Of keep systems, and so forth but. Most of the interfaces, were built point-to-point as. Well it was making our challenge, just a little bit more difficult, as as. We were going there was some data replication that. Was happening across some of the systems that we had to account for and, figure. Out what what to do it, was gonna be expensive as we started to estimate, we actually went through an estimation effort, of what it would take to make changes, all these systems the number was coming up and it was much too big, so. We knew we had an in-state, desire, which. Was get to more of a kind. Of enterprise bus type of capability. -. You. Know you see graph in here that was in our kind, of end state design. Structure, that we were kind of coming up with along. With some, better capability. Of making changes, to these. Changes. To these systems but we had a lot of work ahead of us in order to try to figure out how to get there so. Started. To build a graph alright it said okay if we're gonna do GDP, are we.
Got To have compliance. We got to have employees, we got to have applicant we've, got to have infrastructure. Stuff. From Active Directory etc. And we started to build that and started the modeling, process for, it and, you. Know that first slide I showed you with all of the kind, of interrelationships. We started, to build that not that. Is a small, version of what the real one actually looks like because it's got grown as more. And more parameters. Were coming. Into play but that's where we were thinking we were going to do as we were thinking about building it so. That. Gets us to this, event, one year ago and. We. Were invited to come up we were again, really. Thinking, about how to use it in this context. We were also thinking about ID management. Some other areas, but we haven't pulled the trigger yet and in terms of what we were going to do so. We come to this event and. This. Actually, this this. Is my big plug for this event this event one year ago was really. Really important for us and, saved. Our bacon in some respects, in terms, of, the. Amount of work that we were going to do in different approaches, that we had so. In that, we, came up with other, better integration, layers with. Connectors, out there they're, better discovery, capabilities. Out there. Based. Reporting. Looking. At automation, versus, manual and trying to define some sort of threshold and some, prioritization, so, what. We really found was, there, actually is an. Ecosystem, around. Graph. As, well. And. We started to tap into it and with. That we started, working with focal, point which was a consulting, company that we got real refined in terms of what we needed to do and what we didn't need to do and we, met a company here called clued in Tim. Ward stand up back here Tim's the founder. Of clued in so clued into a company out of Denmark and, they. Use graft technology, and other technology. And they. Had been doing a lot of work around data, and, data management and. They. Were much, further along in that in that. Context. Of solving. EU GDP our problems. In some of the pieces that were kind of wrapped around our problem set that they, had kind of already solved for that. Would help us move along a lot faster so we decided to partner with them and. With. Neo4j, of course we, ran we're running this on Amazon Cloud Services so we could stand it up fast and so forth and so, in the matter of, what. Months I guess really a matter of a couple months after this event all. That structure was kind of done settled, in place and. We're up and running on a project, where we now have a kind, of a. Target. An approach, to, get there so. What we implemented, was, this gdpr. Management. System along with with, clued in Riza. Con here reads a stand up razor hand raises our chief, Enterprise Architect, so. Anything. That got built he's really, the one that built it not me. So. We. Worked with with. Clued in with neo4j with, Amazon, to. Stand up this environment, where we could have connectors, into our application. Set so that we solve this problem not, trying to write code into, every application, we wrote it using. This connector, approach. We. Used the Enterprise Service bus concept. As well, that was kind of, directionally. Where we wanted to head and then. The clued in technology, that they had built was, sitting there and they, had rapper rapper, de round that some UI, to, handle the specific, EU GDP our reporting. And and. Some. Of the things that we're very unique, to EU GDP are and then, at the heart or when then the. The. Main part of the clued in tipple a, system. Is a graph, database where, we're taking all these particular, pieces and, so forth, so. You can imagine lots of ingestion, occurring. When you look across those shared sites across, the SharePoint, sites across the mail and so forth. We, built, the connectors, with the most critical apps and the, apps that we felt had the most data. Associated. With it was potentially. Had. Some sort of compliance. Issue, at hand we. Ranked, every app from, 1:20, and to think, three or four tears, we. Implemented, a certain set of the tears and the rest we said you know what for those will leave manual we'll come back later we'll. Look at building connectors. For those at some point and some are just quite. Frankly they're not worth it right there the the amount of data that's in those applications, would be so small it's, not worth the overhead, of building the connector, we'll just deal with those manually, if we get those but. For the most part everything kind of ran through. This including. Any big systems like HR systems, and our warehouses as, well kind of feeding in through it.
Vision. From a visionary perspective. We wanted to, not think of EU GDP or we wanted to think of GDP, our, global. I showed, that graphic, of us a hundred and fifteen thousand. Employees. Well. The. Philippines', hasn't. Already a GDP, our regulation. That's not quite as stiff as what the EU is in terms of a penalty perspective, Columbia, has issued, GDP, our regulation. What. We see is more and more countries are, probably. Going to be adding, these types of regulations and, in theory. Every. Country can write their own law, all. Right and. We've. Even had one major client. At. A top. Fortune. 50 client, that said guess what we're going to do we're, going to write our own requirements. Layer. It over top of the, GDP, our requirements. The country. Regulations, that are published, and a few little, bells and whistles of our own to that so now not only do you have the possibility, of every country writing, regulation. You. Also have every client, that can put their own spin on it from a contract, perspective. As well that's, a lot of permutations. When you start thinking about, regulation. And how to handle it so our, vision kind, of from the beginning we, knew we had to solve for EU first but we wanted to really think about this in terms of a global perspective not. Just an EU perspective. So. The technology, that we chose and the architecture, that we chose fits well with that we can expand that out as the, regulation. Continues, to go and. Adapt, for those clients overlays as well. So. As we're doing that we're enriching, our data we're adding more will. Add more regional, apps will add you know these tier 2 apps as we need to as they as, regulation. Continues, to grow and over. Time we want to actually take, what are called data Maps which are. Kind. Of a regulatory. Piece, of what you have to do to show that you've got all the stuff mapped we're gonna convert that and put that into, into.
Graph Form as well so we've got a little bit more work to do. So. We've been collecting a bunch of information along. With GDP, are at the same time which we'll get a little bit into where we had it so, we've got five hundred and thirty five thousand, persons, worth of information we. Had collect multiple years. All turnover. Over the years. Twenty-eight, thousand, projects, worth, of data. 2300. Specific. Clients. That's kind of at that line of business level, if, you want to think about it that way and you can see five. Hundred and sixty-seven. Thousand. Worth, of in this case the data points, that we can find interesting relationships. And that's just getting started, because, that doesn't have any operational, results, tied to it that's just information. At, this point which is which is pretty valuable, so. It, kind of gets to where, are we at and. And what do we have on the horizon in, addition. To our GDP our approach. I. Don't. Know if anybody notice, this about Convergys but we're, pending. An acquisition. Synnex. Corporation. And they're. Kind. Of version. Of us customer, care outsourcing, company concentric, is. Currently. In a process, of acquiring converges, so. The last public release that came out for that said it would happen in fourth quarter probably, early fourth quarter that that should close and will be one company, at that point well guess what. 120,000. Employees now becomes, 220,000, employees. Every. One of those data points that we talked about in terms of. Being. Able to graph and be able to look at relationships, almost, doubled, and. In terms of scope, so the opportunity. For us, is. Is, tremendous. I think in terms of this even. Just looking at integration opportunities in, our other areas where we can kind of lever this just from an integration perspective, on. Our own and kind of combined, with the larger company, we've been looking a lot at data. And how to drive operational, improvement. And operational. Efficiency, so. The data that we've already been, kind of collecting, in this knowledge graph we. Want to start, adding more operational. Results to it as well, think. Of a pick. An operational, result down at a real fine-tuned lay layer.
Let's. Say we have a problem with tardiness. Okay. In a region of the world being. Able to associate, tardiness. Results. The people the. Supervisors. The clients. The sites the, region's the management. -, how, are operating, results for that client in general, what are the overall trends, for that client in terms of tardiness etc. I'm being able to kind of merge that operational, data and to, drive some. Operational, analytics that, can, also be performance. Based. Operational. Results as well so we're looking to take and expand, these. Data sources that we have in order to drive more operational, analytics we. Do a lot of our PA work robotics. Process, automation, where we're trying to drive, through. Through our PA. Improvements. In a client's operations. Well a lot. Of that discovery. Work. Associated. With our PA is manual, it's, based upon. Knowledge. That, particular, operations, people have or that, IT, analysts, can go in and be able to find and so forth in order to operate to. Initiate. What might be operational, improvements, what, we want to get is more and more of this operational, data to where we can have points, to go in and look at and say, these are areas where we should look at for our PA. And have. More of that kind, of pre discovery, work automated, so that's another area for us log, analysis. We've got tons of log analysis, that we're not getting any value out of or. Very limited value we think that's another kind, of opportunity, area then. Some others you know we're doing work around smart, help making good, automated decisions, off of chat. Natural. Language and m/l kind, of proof-of-concept, activities. Going on biometrics. Especially, facial, biometrics. And security, solutions, in general so you'll see kind of I. Tried. To list a few of the things I thought connected, data were really kind of tied to. Large. Data loading, we're seeing some cases, where 10x, improvement over.
Previous. Kind. Of models, that we were looking at so that's an area where we we, think we can get real good value, we talked about log analysis. Employee. Improving. Our knowledge graph one, area where the knowledge graph is that, we got an interesting use case right now that we're kind, of doubling down on is. What. Do we really know about our clients. So. We've, got all of these data points around the company especially a large, customer, that's operating on a global basis, and so forth and the. Way we actually get. Knowledge about our clients, is pretty. Archaic really when you think about it we. Only we, have a SharePoint, site to have all these different data elements about a client, and if we get new when we add another column. In the SharePoint, site if you want to think about we use salesforce.com. Which. You, know is, interesting. But not like, super great in terms of being able to, do. You, know make discovery, off of information, around clients, and so forth so, we think there's a value of, information, floating. Around the company that we can create some better leverage in terms of that information about clients so that's another area kind, of long the knowledge graph concept. That we're looking at to. To. To. Improve. So. Picture-wise. We. Think all of this can kind of relate to each other this is, Reese's. Favorite graph. Where. We can take the data, and we have a lot of data within the company, how we can do statistical, analysis, and reference, off of that how, we can tie. That in to machine, learning how we can tie that in to our PA how, we can tie that back into. Omni. Type, solutions. That could help benefit our customers or, at some point benefit, our internal, operations, as well we're. Just getting started. But. It's an interesting journey that, we've kind of went through at, this point led. By. Trying. To figure out how to solve GD P R. So. With that that's, all the material, I have. Welcome. Any questions. Discussion. Yes. That. We. Kind of went in early on and, said. We've. Got this big discovery, problem that, we've got to come out with and. We actually this, is an odd combination, we. IT partnered. With legal. Alright. I, usually. Try to stay away from legal, right but we partner with legal, and said okay if we're gonna go in and we've got to do this discovery, because in some cases you're going into SharePoint, sites, you're going into shared drives you're trying to dissect, business. Processes, that are, at the operational, layer and you're, right by your question people don't really like you to do that and so forth so we got enough momentum kind. Of started, at the beginning that, we didn't get a whole lot of resistance. Some. But, not not a whole lot. Any. Other questions I. Would. Just encourage you all on. A parting, point. I've. Been to a number of the different sessions around the day and there's there's a common, theme in every one of these discussions, and it's, that most of us are. At. A point, where we have a good use case or we're. Trying to, fine tune a, good use case, to. Figure, out how to use it. Pilots. In this. Infrastructure. Are easy to stand up very.
Easy And. Not. Cost prohibitive. Either so. It's easy to kind of take something, run with it and in, my opinion if. You didn't didn't, go where you wanted it to go. Scrap. It and figure, out another use case that might work for it and so forth we've got a number of pilots underway I think. A couple of those are going to materialize pretty, well. Maybe. Not all of them but, pilots, are easy to do, it's. Pretty easy to get started it started skill set wise around. These projects, as well, I'll just encourage everyone to do that yes. Of. The graph it's the. Graph model itself the. Graph itself. The, graph, itself, that we're using for. Gdpr. Is, included, okay. That graph is sitting include, in now, tim is going to be speaking here, at right. After I'm done I think a four o'clock over in one of the larger, rooms I would, encourage you to go to his session, because, he's going to actually dig into some. Ways that, they're doing this at. An underlying level, so. I would encourage you to go to that session, as well if you've got interest, and I think you'll get a lot, more, detail, than I could possibly even answer you kind of from that view, okay. All. Right well thank you appreciate it.
2019-02-08