 Welcome back everybody and welcome to the next session on enterprise knowledge graph in action at LGT My name is Tony Shaw. We met in the introduction. So Pleasure to have you back a couple of things before we get going Everybody in the session except for the speakers, of course are needed, but you can leave in the Q&A section on the right hand side and We've agreed with the speakers that will handle most of the questions at the very end. So Please be patient for that. They have allowed specific time to answer questions at the end if you're finding that you need to enlarge the screen at all there's an icon on the bottom right of the the Screen image which will enlarge it quite significantly. So take advantage of that All right So let's get started our two speakers today are Rhea Wink the head of data analytics and information management at LGT and Jacobus Gillop the CEO of agnus.ai I'm pleased that we were able to squeeze this case study in late in the planning process because it is a significant story and I'm grateful for both of them for bringing it to us so Rhea and Jacobus, please take over Thank you Rhea, you take off Okay, then I Appreciate to being here We will talk in the next 30 minutes about the enterprise knowledge graph in action at LGT We will go through shortly an intro Just to set the scene and then we deep dive into what did we deliver in this EKG project together with agnus.ai I We will show you how we built the EKG up at LGT and then do some deep dives on the key topics Going through the success factors for the EKG Implementation and last but not least we want to give you a bit of insights what we learned on this journey All right, just I Will start as Tony said I'm the head of data analytics and information management at LGT in this role I am responsible for the group by data management and analytics practice is the LGT This is covering defining the vision the strategy road map Building up architecture infrastructure and also the delivery of DMA Data management and analytics services and products and in this role I can count on a lot of committed and skilled team members who are Delivering their products and solution before I had the opportunity to get in this role I was in the Institute for Information Science and did their research projects Mainly focusing on semantic technologies I hand over to you Jacobus. All right. I am Jacobus Gullurk. That's a Dutch name I worked since 2010 on semantic technologies working for companies like JP Morgan Chase and Broomberg and B&Y Mellon where I worked on semantic technology projects we started to call it enterprise knowledge graph at some point in 2015 and Got a large and the EKG as we call it now in production in 2016 It's still running in production serving regulatory purposes and many other use cases And after doing it a few times I thought let's start a company doing this Focusing on just this this particular topic In 2018 and we started talking to LGT and worked for LGT ever since All right Ria, what is LGT? Thanks, I just want to introduce LGT so that you get a bit of an understanding in which environment We build up the EKG LGT is the world's largest private banking and asset management group owned by a single family The Prince the house of Liechtenstein We are delivering sophisticated private banking and asset management solution Covering traditional private banking services whilst planning lending and financing Alternative investing and even investing with impact as you can see LGT is present in more than 20 locations all over the world and We have about 240 billion in assets under management and around 3800 employees at the time Just reshare your application window, and then we'll post it again This I clicked on the wrong Chrome tab. Sorry for that. Okay It's all right Coming up loading and now needs to share this again share screen Chrome tab And there we go Apologies for that Perfect so We want to show you a bit what were the key achievements of this Project or initiative There are mainly six main deliverables we Developed during I would say one and a half year Important is one deliverable the KG strategy and architect One of the key topics we will also talk about later on We developed an EKG strategy, which was strongly aligned with our business strategy and And I think I can hunt over to you to co-boost explain a bit the extensible EKG platform Yes, yeah, what is an EKG in the first place, right? That's that's our name for Enterprise knowledge graph obviously, but it's it's it's a knowledge graph at enterprise scale. It's designed from the ground up to support infinite scalability Allowing you to build a digital twin basically or multiple digital twins, so we will for creating a model of the real world getting all the data in your Space that could be the whole bank or even a beyond the bank Getting all your data sources on board it linking all the data together without necessarily Putting all your data in one database that data could stay in place Or it could be moved to special types of databases graph databases For example, triple stores cement semantic graph databases at LGT. We're using starrock for that And We are not saying that we would like to have all your data in one in one database and we are also supporting The principles the 10 principles of the enterprise knowledge graph foundation Which basically set out these are the 10 things that you need to have in place in order to to create an EKG and We'll talk a little bit more about that. That's one very important concept is The ability to support Open world basically support all the most granular data of all the various sources next to each other And and present it all together link it all together So the exact lamp base Yes, go ahead. So based on this Extensible EKG platform. We'll build up our EKG lighthouse project. So the first project we did in LGT based on this platform Which was was talking about the EKG lighthouse project was quite a special one We developed a legal entity management solution and we replaced the existing solution by doing so We tried to find a use case Which has high reuse potential So that we also show what are the benefits of EKG for the entire company and not only for a specific use case And we're trying to find the Project which is from the complexity side not too complex so that we can deliver it all fast But still that it has enough complexity to show as I said the different Benefits of EKG as within the data management and analytics platform Exactly and next to that. We also delivered basically the first Yeah cloud based development practice like the cloud came basically at the same time as when we started. So This is kind of the yeah a new complex kubernetes based open shift based Deployment many different services and environments Running at the same time. It's all infrastructure as code DevOps We're using a team of let's say an international team that uses github and slack etc And that works very nicely now with an internal team where so basically our test and dev environments are in In the cloud and and we we we work together in a way that that that works well actually and which is kind of a A nice model to scale up resources as well for for for lgt exactly and Also Yeah, we are jumping for the back with um develop the method for dma projects Which are based on enterprise knowledge graph Just it's a standard space. So we didn't invent the wheel. Um, we just took what is there and tried to Implement it within lgt um It is really covering end-to-end from how can you analyze the strategy? to really implementing The use cases on our infrastructure and it's also aligned with the ekg method from and the majority model, um, which cacobus talked about earlier I think that's quite an important deliverable as well because it is different to build up ekg projects within a company and therefore It's required that everyone in this project knows what to do and how to do it Yes, so so we have a team of people where some of them have worked In multiple ekg projects already Uh, so there's a lot of experience many years of experience now that we kind of capturing now in what we call the methods The ekg method and we are working with the foundation to make that a extended and open or public standard But part of that method is the maturity model, which is kind of a a structured way of translating vision and Into strategy and creating all the cake basically addressing all the capabilities that are affected by uh by ekg Um, and that it's it's a structure where we basically say we have four capability domains business to start with then data then technology And then the rest of the organization Um, so that's the four pillars as we call it usually, uh, not capability domains, but that's what they are We try to address basically those four different audiences in their own language as well Um, and then we have five maturity levels. Uh, so you start at level one with a one use case for example, and then the platform level and the enterprise level Um, and if you want to know more about the maturity model, then please, uh, yeah, please join the enterprise knowledge graph foundation I would say where we are working this out and capturing this all in in in the method Exactly, and there's a second deliverable we achieved within our ekg project is the center of excellence or center of competence Doesn't matter how we call it. Um for us, it was important that we have Um, really in-house capabilities around the ekg. Um, but still we have a balanced sourcing model, which means we are also relying heavily on agnus ai and their expertise I think it's quite important where you will never have as a company from scratch on all the capabilities you need to build up an ekg Which is um, really a basis to use it for the entire company Um, we have also evolved the ekg capabilities within lg t through trainings and further developments And um, I think that's that's something We have learned and we will go through it also later on That a balanced approach between having in-house capabilities and really a strategic partnership Is um Crucial to be able to deliver fast and and to plan also ahead the roadmaps how we will Expand to use cases is backs Based on the extensible ekg platform. We build up at lg t and as you can see, I mean these are six Very different deliverables from from strategy from method over actually really delivering added value through our first Lighthouse project and having then this this platform which we can use further on for dma activities within lg t I think Now that you have um seen what we did within I would say One and and maybe a year and three months one and a half year Um, I want to show you a bit how we started just going through it Where we started what we did so that you can understand a bit the journey and as you can see it was um already started in 2017 Within lg t we did an analysis of our business and it strategy to understand what are the strategic objectives And how can we support that from our department? We also performed two pocs to proof of concepts one was in the field of artificial intelligence We called it back then cognitive computing and one poc was a poc around ekg, which was basically The groundwork for then really delivering the the lighthouse um project the legal entity management solution And in parallel we did also a readiness assessment for data management and analytics covering also the ekg because in my opinion it is quite important for companies to understand that it's not about artificial intelligence or new technologies. It's often really also About the the culture and the way of working and the methods and the strategy Within a company and therefore I wanted to show in parallel of the pocs on which level lg t has to do their work to really improve by using data and analytics and based on these readiness assessment and the poc We did back then we got the mandate From from the senior management to sea level management to build up a holistic and strategically aligned data management and analytics program within lg t which is As you can see really based upon the ekg and it's covering the governance part It's covering the data factory the data factory is our platform We build up where the ekg is is the core capability within and as you can see we have four different service groups So we don't only deliver the ekg for analytics. We use it also for data management and information management functionalities and for business intelligence or reporting and dashboarding solution as well as for use cases In the field of advanced analytics and artificial intelligence When we have defined the the vision the mission and and this approach how we will institutionalize data management and analytics and the ekg within lg t We looked out for for partners who have the knowledge and experience to build up such a platform Within a company and we reached out to agnus ai and then the the hard work started Because doing a poc is quite easy. I mean it's it's It's a narrow scope, but really implementing a platform and bringing in this change of of working and and also this strategic view on on data management and analytics is quite A lot work to do and we are still Doing it to be honest. It's a journey. It's not a something you can do one off and then it's all good And together with agnus ai We had the ekg vision alignment. So we really put the ekg in the center of the dma strategy We performed as I said before this lighthouse project and followed this balance sourcing strategy um We talked about before and based on that we are now working on a roadmap So how can we leverage the potential we have based on the ekg platform? Maybe we can go to the next slide to copus. All right Thanks We have Chosen two main key topics. We want to really go through Thoroughly in this meeting on one hand the vision and business strategies And on the other hand the lighthouse project or how we choose the the project To start the ekg initiative I will shortly go through the business strategy in my opinion and i'm pretty sure to copus Is Think it as well the business strategy So as dma or ekg responsible person you really have to understand The business strategy what what are is lgt focusing on on what are the strategic objectives? What do we want to reach by building up an ekg within a company? And for us as lgt is clear, we want to stay competitive also in the digital age we still want to provide comprehensive and client focused private banking services And want to provide an outstanding expertise and service quality. So this means for us Who are building up an ekg as supporting technology for the entire data management and analytic services That we need to be flexible and still reliable in what we are doing Business also stated out that we have to increase the agility so That we can react fast on the changing environment We have from business side, but also from a risk and regulatory perspective and as you have seen in the introduction of lgt Ethics and and and data privacy and securities a hot topic in in private banking and lgt specifically So the solutions we are building up on based on the ekg They have to be also Reliable and they have to follow all the regulatory requirements. We have in financial services A point. I think everyone from you hears it Every day We have also to to decrease and I would say drastically decrease the time to market for our dma solutions That we can when there are upcoming new ideas from business sides or pain points that we really can address them fast And that we still can doing it But still can be flexible and also offering Really specific services, especially in private banking. Maybe that's in retail banking a bit another thing But in private banking, it's important that we are client focused and that we can deliver tailor-made services based on the ekg And the first point I think is something It's good that business really point Pointing it out. They saw that we have a bit of lack of an understanding of the data we have So this 360 View on clients or on assets or on processes or services Is quite hard to reach if you have so many legacy systems And therefore this is also one of the main points we had to address by building up the ekg Maybe to cope with you want to add something Yeah, yeah, especially that yeah that 360 degree views that that that goes by many many different names so you can also call it the holistic view or Anything 360 there's all kinds of names for this but basically it means like like if you present a user interface to an end user that is kind of very tailor Very bespoke for that end user Then we would like to offer the ability for that end user to click on anything And and dig deeper basically and find information that you've never seen before Like that you can click around that you can discover data basically That you can see the full picture And get a better understanding of Why do I even see this data? Why do we have? How does it relate to other data and information? So that we can eventually offer better services to customers and with more understanding of the customer Let me switch to the next one. So the data strategy Like I just said is the holistic view basically again making a data-centric approach data is a primary asset maybe the primary asset And and I think yet Ria you would probably want to say something about this like technology Follow the data Exactly that's a discussion we always have within it because our department is actually In the it i'm reporting to to the cio directly And normally you have the business Strategy and then follows the it strategy and maybe then follows the data strategy And that's in my opinion one of the paradigm changes we all we really need to do That first is the business need and the business strategy And then we have to align the data strategy and the technology follows to support the business and data strategy But this is really a change which Is hard to do to be honest and will also take time I would say years to really get this mindset into the sea level But also in each and everyone who is working within lgt um From recent side as well as from it side Yes, and we do it at all levels. So that this is this is very helicopter level But when you zoom in on one use case then it still is that order like business owns the use case Defines what the use cases is we try to create a so-called use case tree Which is part of the method more a little bit more about that later Basically plotting out all the various use cases that the business Wants us to deliver And for each and every use case we get first business needs and then translated to data strategy to to to data requirements And linking it to things like ontologies and data sets etc And then eventually we worry about the technology the technology is the least of our problems basically It's the technology works and we can fix everything But it's it's like getting things aligned business needs data then technology and then the rest Um, and and also it's about much higher levels of quality Oh across the board, especially data quality Aim much higher. There's so much we can do so much more like all the data at the moment This is in silos as it's biased. It's designed for one purpose Um, whereas in a in a digital twin and in ekg The data that you present in the in the ekg is kind of By definition is not biased is is is not designed for just one use case. It's it's mapped to all kinds of ontologies Could be conflicting ontology dealing with multiple versions of truth at all levels For example for one particular Uh Jurisdiction you might have this terminology You could have regulated that calls a critical business function Something else like it could be a critical operation for example In the uk they call it critical business function. I believe it in the u.s. They call it critical operation Same thing, but you have to deal with all these different models different But it's not just different terms, but slightly different ways of looking at the world And in in an ekg your data has to be so fine grained and and so Normalized in that sense that you can deal with all these different models then Next to that it's not just working with drum data But data we want data with machine readable definition So that you and and link to knowledge that you capture in ontologies, which is basically your machine readable knowledge Uh linked to all the types of metadata that you have that's not just the description of the term But there's much more to this is the provenance and the and the lineage Who owns it and what systems created it and aggregated it? What are the The let's say the retention policies the caching policies the pricing policies You name it. There are tons of different types of metadata that you all want to bring together Then on top of that we want to also know what's being done with that with that data, which is what are the use cases? How is it used? Who is using it? All of that comes together in one place And that's that is your data strategy basically and we link it together By having this core artifact that we call the use case tree Which is an artifact that everyone can talk to in all those four pillars the business people They know, okay, this is the use case that we that we are working on The data people know it What what it means and of course they have all kinds of translations to their own world and their own lingo But the core artifact that we have from we get from star to finish all the way to production and and beyond Is that use case tree? So and let's let me switch to the To the oh, yeah, so one important point and for Koti it is the open world story So how can you build a digital twin of your whole world if there's only one single version of the truth? That's not a digital twin And that is the censored version of the digital of the of the world maybe but the real world is There's always multiple versions of anything Um, so what we like we are doing we're changing that that's unique the unique feature Let's say of what a real ekg is is that you embrace reality That you institutionalize the paradigm of open world that you basically get all versions of the truth next to each other presented And then let's say per use case You uh, you come up with the right version of the truth for that particular context. Everything is about context Um, then next to that we have the enable the discovery already mentioned that in the previous slide So what's not in now and let me switch to the next one The technology that translate all of that to the technology strategy Well, if you want to have one platform that runs all your data ultimately then you have to do that in the clouds Let's not talk too much about that. That's an that's an obvious one. I would say Um, so you could call that ekg platform A data fabric or semantic data fabric or digital twins kind of all the same And uh, what what basically happens is because you have one platform that serves all these different use cases all these different product owners and users with and they all have different agendas and they're all important, of course So you cannot work with it's just impossible to do that with standard waterfall like where you You cannot freeze you cannot have a code freeze anymore. You cannot It doesn't work like that. It's like we have to basically do the same model as Netflix and facebook and google and air bnb are using those systems are never going down Um, and they gradually change the system in very small increments and which you could call continuous deployment Which is a key a key feature that you have an in a in a bank in any large bank a large organization an even smaller ones like lgt Continuous deployment is a very hard thing to to to get to get working It's because there's some there's all kinds of checks and controls and tests And user acceptance test procedures, etc And all of that needs to be looked at if you want to go to a continuous deployment. So that that's a big thing to think about And last but not least like reuse is the number one priority in all its aspects And the many projects reuse is never on the agenda even Let let alone a priority But whatever you do in ekg Everything we build is built for reuse and for usability if it's not if it's not reusable then don't do it basically Yeah, exactly. Maybe I can jump in here because that's that's really important as you have seen now also on the slide how we approach The data management and analytics part within lgt you have seen that we have to deliver from various service group from data management over bi to advanced analytics and if you don't Have this view on what component can I reuse? It's quite hard, especially such a small team we had in the beginning To really build something up, which then can be Scaled up and that's also the last point we have here from an architectural point of view I think reusability can only be supported if also the architecture behind it is is built like that and In our also within lgt We have an own architecture department still the data architecture role is when within my department And this is so critical if you don't focus on on the reusability You're just building up new silos and therefore it's important that you have this alignment from from the strategy From a data-driven architecture to really delivering the the single use cases But still based on a on on the principle of of the reusability All right. Well, we have eight more minutes for our part. So let's There's a little bit Faster, but yeah, like strong leadership skin in the game Like rio. I think you can maybe talk to this better than I do okay Yeah, the strong leadership. I think it's important that you have someone who is really Seeing the benefit of what we are doing and and and really Leads the entire discussion because it's quite a lot about stakeholder management. So As chikobu said, it's really you have to have someone who has the skin in the game and everyone in the team has to To deliver as well And this is quite important to communicate always the vision the strategy Why are you doing doing these things? What is the benefit of it? So that you really create the buy-in also from business side as well as from management side And as we already have said the strong collaboration. I think that's what that's also A key from organizational perspective that you know, which capabilities do you need in-house? And you want to build it up which capabilities you need from externals and have a good mix between externals and internals Just to be able to to deliver all these Components and this project so for us as lgt. It was clear from the beginning We will have an ecosystem together with external partners like agnose AI But we want to have for the entire ekg at the end of a journey all the capabilities needed within lgt good let's jump from the discussion about the the vision and strategy alignment which is quite the core topic if you are building up ekg in a company Which is want to give you some insights how we choose our lighthouse project. So the one project which should be Giving us a bit more of support also from from management level. So for us it was Important to have a project which adds really value to the business so There must be a clear pain or need from business side, which you can solve by applying ekg within such a process It's also important that you can show that it's not only for this use case So the problems you're addressing within this first project you're doing based on an ekg should also have actual problems which are common in a in a company so that you can show That it's a ekg not only a silo solution, but really something you can address so many different use cases within a company Yes, and yeah, we we also we started with basically a three-day workshop where we Discussed what is ekg? What are the business benefits? How does it affect the business? Etc. But also what we call discovery and like this basically think broad think big What are the long-term use cases that you would like to implement and then drill it down to? particular set of use cases that That that have a high reuse potential and would and basically create a roadmap like a shortest pathway To that high-level roadmap a use case at the top of your use case tree that you would like to implement So in this case it it's all about basically the core of the bank all the other reference data Start with legal entities. Let's say and contracts and and all the various reference data elements around that and I think Spending some time and thinking thinking broader and thinking about high reuse potential is is very important All right, let me switch to the next slide because we have only a few minutes left all right Exactly, so we already talked a lot about that. What are the success factories in building up an ekg? We split it up in the pillars to cope as mentioned before from the ekg maturity model Regarding the business pillar. I think without c-level management support. You cannot really be successful because it takes time It is an innovation project. So Not everything will go in the right direction So you may have some setbacks and some learning curves you have to go through So it is really crucial to have someone in the c-level who is supporting and understanding what you are doing so having really this clear mandate, but also the the leeway and as we already mentioned in the beginning having this shared vision what you want to reach with an ekg within within a company and what its value is something we may miss also during the entire project to pronounce so from from time to time And last but not least having an understanding what is the scope of the first project Is is for sure helpful if you don't have a two broad scope and you cannot really deliver in time I think that's also important that you have there the support from the business owner So our business owner of the ekg solution we build up with the lighthouse project. He was really passionate and supported our Initiative, I think that's quite a crucial factor in being successful in delivering ekg projects Yes, I can only See of this with this of course like the the c-level support is for me the critical one Because I've done multiple projects and the ones that are were successful were the ones that had c-level support If you don't have that if then basically almost don't don't even start If management only wants to do a p or c that's kind of low risk No skin in the game p or c barby where we basically introduce At least 10 different paradigms in one project doing Basically everything in a different way than what people are used to Then the pushback that you get in a in a in a large organization is just too much Too much to cope with And getting it in production is as almost impossible like you have to have top level support You cannot do this bottom up. That would be basically my if you remember One thing about this presentation. That's what I would say. That's the that's the takeaway And it's about mandate and leeway and everything else derives from that basically Let's go to the next one because we're almost Exactly as I said before I think in the beginning we were communicating in the company quite a lot of what we are doing We get got some insights to the different Stakeholders, but during the project we may lost a bit track because we were focusing so On the delivery that we didn't have time to always explain what we are doing and always Give to the the various stakeholders. We had the insight they needed. So I think that's something we would Or should do better Next time that we really over and over again repeat. What is the benefit? What are we doing? How can it evolve in in LGT's Portfolio, so I think there we may lost a bit track over the over the project What we saw also saw and the point we mentioned before is the architecture So We had quite a lot of discussion around design principles and architecture principles When building up the platform In the use case so also there I think That was mainly because we didn't had a strong architecture When we when we started I also think that the distinction between the efforts we we have to build up the ekg and the efforts we have to build up just this lighthouse project That's something we may missed a bit. So we Financed everything the ekg platform and the lighthouse project through a business project And this also put a bit of pressure to be able to deliver So I think there We will improve the next time That we also be a bit more transparent about what is the ekg platform effort and what is really the effort Just for the use case we have to build up. So yeah Jacobus and ray we have quite a few questions. Would you be willing to switch over and ah, look at that? Right on time A question that's been asked in a couple of different forms is can you share what a typical business problem Example would be that the ekg will address effectively Well, at the at the moment we what we delivered at the moment is is Legal entity management like said all the the whole workflow Around creating new legal entities for for example a front You create a new legal entity that is owned by multiple parties and you have all the various details and Differences in jurisdictions and Directorships ownership control structure Like all the various details need to be Approved etc. We have auditors and accountants involved. So there's a workflow around it with lots of details So getting that data to to be correct and precise and high quality That is in this case What the use case is but but you could you could Extrapolate that to to many other use cases like in any given organization, especially in the bank All all that all that master data reference data needs to be needs to be of high quality and I would say we can we can We actually replace the system in this case But I'm not saying that ekg is there to replace your current systems Which you can see there's a layer on top of all your other systems providing a holistic view And and figuring out what is the right data without doing that in etl And in five-page sql statements anymore But do that in the context of a graph and creating this high precision high quality Layer of data that can then serve many higher level use cases like risk management Supply chain management like there's all kinds of high level use cases, which we call the enterprise strategic enterprise level use cases that can then be built And that's I And that's I we have this bonus slide there the what what next Is basically that's what we're trying to do in the next for the next series of use cases showing that we can decrease the time to market that we can Deliver much faster and faster because we reuse lots of underlying use cases And so everything we build so far is 100 reusable and completely model driven There's no single line of code for this single for this first use case Okay, so uh, I'm gonna I'm gonna extend that question a little bit because there's a specific question here from christen who asks Would the the um knowledge graph? Replace the entire warehousing or aggregation strategy? Tricky question Very long term Yeah, exactly exactly. I mean Let's say from a recent perspective. Yes, it could and should go there The time to do that is uh, I think we are talking about years And that's actually also why we had in our platform not only the ekg, but also traditional um, let's say data warehousing data analytics technologies Because putting everything just on the ekg would be way too much But over the time, I think it's it's going in this direction for sure Okay um Our friend mike Bennett is curious, uh, how do you ensure that semantics? The semantics or meaning is owned by the business rather than the data side Mike is on the call mr. Fibo Can I can I take this one of you Yeah, of course, of course So What we are trying to do here is not Presenting ontology into business people. Uh, I've I've I don't believe that that works. So we we came up with this Yeah, new methodology, uh, that's around use case three the use case three concept Where we define, uh, what the business outcomes are for each use case Uh, and eventually also high-level user stories and like for each user story, let's say feature or Requirement as a persona. I want this feature in order to achieve this business outcome That is the kind of language that we can, uh, get agreement with with the business Without talking basically stripping all discussions about, uh systems and and screen designs and database schemas or ontologies Let's not talk about that. Just tell me in plain english What do you want? What do you want us to deliver to you? That's uh, and that's what the use case three is all about and then Don't walk away and translate that to some other model. Uh, that no one recognizes anymore But keep that use case three alive and in tact just add more information to it Which we did and by adding all that information eventually That is your use case that is the program. There is no coding. There's you just add more information about What is these user stories? How does it work? Uh, what kind of data are we combining? Uh, from which data sets, etc That is the program. So that that's what I mean with a no code, uh use case. There is there is no code It's all modeled with them. So your use case three eventually it's like a christmas tree you hang more more Stuff in your in your christmas tree and eventually it runs in production And maybe I can add there a bit, uh, let's say more from an organizational perspective because Actually, I am when I'm talking about data I think it's Important to see that it's more a business topic than it is a classical it topic and We are quite happy to have subject matter experts from business who are heavily involved In this kind of data steward community and because they have issues With low data quality. They are also contributing a bit in in understanding better. What what is the data within all the applications and Getting a better understanding Of the meaning so we are trying there or with this data framework data governance stream to engage business and I think we are on a good way because Data and digital is also in the business strategy of of LTT So they acknowledge the the importance of of really managing and using data properly Uh, Jeffrey has a lot of fairly detailed questions here, which I'm not sure we have time to get to so I'm going to apologize to him for that A couple of quick questions though, uh, can you comment on what kinds of volumes? Be it customer account arrangements That will be handled through the knowledge graph Um Point is currently is only the the um lighthouse project is in Production we are working on other projects, which is expanding the the lighthouse project as well There we are talking about different Volumes to be honest because this lighthouse project wasn't really about handling volume This will be a topic we will have to discuss and to address In the further roadmap and and use cases, but it wasn't meant like really Having to control over a lot of data. It was more meant to really being able to manage the legal entity Uh data and metadata and so on so it's Currently not not the Main topic It's a use case that doesn't have high volumes, but uh, just to say That doesn't mean that we can't handle high volumes. Actually At the previous company that I worked for uh, we were dealing with Like data source that had two one data source had 200 million records That we need to the onboard onboard translating that to a billion triples. Uh, that's just one data set So the the the triple store technology that we use for Let's say behind the behind the the service layer of ekg this there are triple stores That could be as as many as you want and not necessarily just triple source could also be other types of databases Uh, but those triple stores like stardor for example can handle uh, at least 50 billion triples in one database. Um So these these that technology is actually Expanding rapidly. There's all kinds of other vendors out there or other products And there's a lot of competition and and the volumes and the performance Is getting better and better and better I'm gonna have to wrap us up there. Uh, I'm sorry. I I um would love to get into some more detail with you But we are at our time So I want to thank uh rayon jacobus for a wonderful presentation. Um, we'll be taking a short break now In the meantime, we encourage you to network with the speakers and other attendees within the spot me app I particularly suggest rodney and jeffrey, uh, whose questions were unanswered Perhaps reach out to jacobus directly to get answers to your questions there Uh, we'll be back at the top of the hour, which is about uh, eight and a half minutes from now at 10 a.m Pacific time 1 p.m. Eastern. We look forward to seeing you then. Thanks very much again everybody. Bye. Bye. Thank you. Thank you too