Picture This: How AIG Visualizes Propagation of Risk – Deborah Baron & Jim Barrett
My. Name is deborah baron and I'll, be. Introducing. Today AIG. And their. Story. We're. In the final stretch of the day getting. The happy hour so, I really appreciate, all of you being here today. What, you're about to see from. AIG, what, they've designed, I think is gonna make this one of the best sessions, of the day so text. Your friends tell, them to come, and. The reason why I think it's remarkable. And, memorable, is. Because, what they have done what, they have done is similar, to what Zillow, has done, for. The real estate industry, right. They've taken us from looking. At the Sunday paper and paging through real estate sections. And talking to people right. To. Zillow, right. Where, you can get all the information in, one place look. At it from different ways see, how it's related get. You know get images. Get, text, get numbers, it. Makes it easy for the end user, it's complicated, array, of data but. Any consumer, can pick that up and do in fact tell, me in this room who has used Zillow, before. Yeah. Would, you ever go back to using the Sunday paper. Yeah. Nobody right why would you right. Because with, Zillow, not only, do you have all that information in, one place. But. You also have the ability to look upstream, at historical. Data you can see property, values you can see upgrades. And changes to the property, and how that is impacted. Property. Value, you, can see comps, related. Properties, right, if. You're a seller, if you're, a buyer, you can, measure the impact of changes you're gonna make to your property, potentially. You can measure the downstream, impact, of those changes on, the value, of your property right. So this is an apt analogy for. What Jim is going to talk about today Jim Barrett from AIG, and the, smart people in the innovations, team there, it's. An apt analogy because it's it's, it's. A phenomenal. It's a huge, leap forward. From. The Sunday paper to. Zillow right it's leapin efficiency, it's, a leap in visibility, that, speeds. Transactions. Because, of better insights, all the data is there right. But now we. Have the insights we have the accessibility, we have the visibility so. Let's bring that back to organizations. Right. Organizations. Today are struggling, to make that leap right. From, paper, to plastic, right. But they're their burden under the weight of, huge. Amounts of information. You. Know rows and columns of data text. Documents, that have been digitized, but. Basically this mass that is just impenetrable, by, mere humans on top. Of that there are static, diagrams, so there are visualizations. But. They're not interactive. And the moment, you print them they're obsolete and, then. Finally, all of this, is all, this information, is sunk into a, complex. Network, of legacy. Systems, legacy. Systems with information, that flows from one, to the next so that a small impact, upstream. Could, have a huge impact downstream. And that. Impact could have upside, potential. It can also have, upside. Risk the issues you cannot, see you. Cannot see the impact you cannot see. Those interconnected. Components. We. Lack that visibility. And so what I when I was. Thinking about the, the keynote this morning and, amia was talking with us about some. Graph stories right, the medical insurance company, and Adobe's. Community, those, were those. Were problems. That could be advanced. And solved. By. Visualizing, the network right. These are graphs stories, and so when I heard about what AIG, is doing with the integrated, operating, environment, and, that a sexy, name, what. What they're doing with the IOE. Integrated. Object, environment, excuse, me, what, they're doing with the iog to me was I was one of the best graph stories, I think I've ever heard so. It was really an honor Tom. Sawyer software, to to, be brought in by the. The innovations. Group by Jim. - to, work with the design and build the application. And. See. There's so many things so many buttons to push up here I have to make sure I push the right one and you, know many of you in the graph community, have probably, already heard of Tom Sawyer software, so. I won't I won't spam but a couple of minutes some, of you might have even met Brendan, Madden our founder, he's been in the graph space, for over 20 years when.
He Got out of grad school worked. In large network, graphs. At. TJ Watson and then started, Tom Sawyer software, many of his peers. Out. Of academia, have gone on to become PhDs. But in our world we brought in a lot of math, and computer science majors, engineers. And, PhDs, that are floating around the world helping, us, continue. To build better braf, technology. And work with customers like like. AIG. Because. They they think of things that we've never thought of and we collaborate, and it's just a lot of fun so it's really my pleasure to be here today to talk about it. The. The kinds of things that we do are very similar to many of the use cases that email. Talked about this morning. I mean financial, services, things. From anti-fraud, to, catching, the bad guys by, peeling back the many, layers to see who's behind all of it the. The. Solutions, in you know in networks and operating, and environment. And infrastructure, some of interesting, new areas, in terms of architectures, and models and so. At, the base is a platform, to build applications. Like the one you're going to see here in a moment. In addition, there are some exciting new, areas. That we're applying graft, technology. In use cases that weren't mentioned this morning things, like model-based, systems engineering. Where. Customers, like Airbus, are mapping, an entire Airbus. 320. So, that architects, can do design work and maintenance engineers, can do upgrade so that she could zoom. Into one system. And then a subsystem, and then to the individual, components. Because a supplier. In, China, has just changed, the component, they're subsetting it, unsetting it and you're, wondering where this component, exists, across the, aircraft, and we're along that assembly. Line it's. Actually placed so, that you can do change management so these are the kinds of program. Management activities. That are going on that that. I think ultimately, are gonna take, Graff to that next level, so. Just to kind of pull back give you that big picture it all starts with the data and, the, attributes, and Jim will talk more about how that plays into his model name and the demonstration, he's gonna show you all, of, that wraps up into a nice data model, that. Information. Is then processed. Filters. Algorithms. Analytics. To. Produce different views, and this is where graph is really really. Different and valuable, so it's that same corpus, of data right, but different business units have different have different needs they're asking, different questions they're, solving, different problems and, so, we're getting a lot more value, out of that same core set of data it's. An exciting, time I think to be in the graph world so, with that I'd like to turn it over to Jim Barrett, aig. Thanks. Deborah. There's. A few times in, your career, when. You, fortunately. Get to work with Superstars, and. Building. The integrated, object environment, ioe POC. Was one such time. Gordon. Cooper planted the seed Shaun, Keenan, developed, it nurtured, it grew it and in fact if you read Shawn's book financial, innovation. It. A financial, institutional, advantage, you'll, see a lot of the a lot of the the, core principles, of the ioe they're gene Goffman, who designed the data model for AIG that runs on neo4j, brilliant. Data model very flexible. In, dubrio who loaded neo4j, with, our with. Our information jason, massimo who put the topology. Together. From, dozens. And dozens of, diagrams, done over the years I. Also want to thank a I geez model risk management, team countless. Names I can't. Even begin to. Mentioned one name and then you slight somebody who didn't mention the erm leadership, team all. The way up to the chief risk officer, and. The AIG IT team. Along. With all of neo4j, z' experts, Thank, You Daryl and and all of his colleagues, and a, special thanks to Chris, Tanner of Tom, sir software, excellent. So. This. Is amazing that I'm actually standing, up here.
Not. Far from where Superman, and, the amazing Hoke are but. I'm talking about something, that's been, near, and dear to me since, I was a boy and this is an airline, timetable, route, map that. I collected, when I was about, yay high. And, I was fascinated, with with. The, visualization of the potential, now I obviously, at age 9 I didn't understand, that but. Now after, studying the, the, the, power of visualization over, text and tables I can, appreciate what I appreciated at 9:00 which was this, in a single, page is, the potentiality. In this. Case of mobility, and so. That's been my theme, through, my career, starting. When I got out of college in 1980, I began. In earnest to do diagrams, in, a way to help myself understand, material and a way to help the people that I was working for as a business analyst understand. Their, work role. So. At, AIG we've, had a, few, problems to solve know when, when we realized. The, potential of what we were doing we went full. Tilt all the way to step three let's. Make full, use of this let's make use of the. Attributes, as we connect them to the objects, and the topology let's. Do some really sexy stuff and we did and we did but. Lately. The, business leaders of AIG, have approached. Us and said well you, know our big problem, is we've been asked, to draw. A process, maps for, the whole company each unit, at a time and, we. Want to we. Want to associate, those process, maps with, particularly. That. Hold down that, tell us about those, processes, that's, exactly, what step one is a repository, for topology, that's what the ioe does first. And best. Secondly. Where's, the risk and where does it go so I think of risk as failure. At a point which. Would be mostly your IT failures, failure of an application database, and failure. Through risk propagation, and I'll talk both of those two both of those in the demo and then measuring path and tree risk that's, perhaps the most complicated thing we'll talk a little bit about that demo. So. We. Really started on a relational, database years, ago, storing. Our our diagrams. Or into structure diagrams and we. Went to the vendor and we said well we need to do some pretty complex, queries, now. Complex. Queries, we need to go 20, hops back on a hundred nodes and we need to know some, particulars, about the topology. They, came back and they said well nine hops is about the most we can do in a reasonable amount of time after, 10 hops we kind of we kind of fall over so. That, made. Us realize either we need a new technique, or we need a new technology, that. Led us to neo4j. Best-in-class. And, they are adding powerful, features. Ketel. To be I mean I can't, say enough about that we're excited about that Seifer, is the language I, just heard last night at dinner that the. Core parts, of cypher, will be likely. Adopted, by the ISO, to be, the GQ, gql. Which, would be the standard graph. Database, query. Language and then the support I want to thank Darryl and and. All of his colleagues for. Support. Across AIG. And. As, well as a great deal of patience. So. After. Neo4j, was. Found I went to the neo4j website, and I looked at well who's who, the visualization, partners, and I looked all fine companies, but, one stuck out in particular and that was Tom Sawyer's software, because of their core competency, is exactly, what we needed the, data flow diagrams, we were looking at they. Had been working on for a long time so. I called up Brendan and I said Brendan this, is what we're doing here. We're interested, but. Our number one focus is, is. Your, visualization. Intuitive. Is it, fun is it easy is it a polite quality, because, that's what we need because, in something like the ioe as you may have already started, to think about the, danger is it. Gets stale the information, gets stale and as Deborah, said you. Do a Visio diagram, and and within. Moments it's it's out of date that's, a problem no matter where you are but, a particular, problem when you're advertising to, the world that hey I've got the latest and greatest on your topology, so. My. Number one way, to mitigate that risk is, let's get something people want to use not have to use but want to use Brendon said hey. We've been working along those same lines for two decades, so, that, cemented, that relationship. I might, also want to add the enthusiasm. And professionalism, of the, Tom horse Tom Sawyer software folks I'm. Overjoyed. To be working with them. So. This is it this these are four components, of the integrated object, environment, it's, a patent-pending, tom patent-pending. Concept. By AIG. Using. Two, technology, components, a graph database and a. Visualization, layer, so. You have Tom Sawyer perspectives, on the visualization you have neo4j, and then, the two content, components, you have diagrams, and, you have attributes, I'll talk a little bit about diagrams, in the next slide the, attributes, you know it's a little bit of terminology so, I know, that.
Neo4j, Uses, the word, properties. All, fine and good we will use attributes some people use reference. Data or the technical term structural, metadata. Characteristics. It's all it's all the same so, think of it as properties, or attributes. That's. That's, a key that's a key ingredient here, because as you'll see in the demo those. Attributes are used to. To, run queries to, go to different dimensions to. Actually. Even be used in, algorithms. And eventually optimization. So. About diagramming, and because. Visualization. Of processes. Procedures and infrastructures, is near and dear to my heart. It's, my favorite topic and I'll say that in my tenure. In. Whatever field, I was in with the airline industry or. Whether. I was in financial. Services and, the brokerage, business or, insurance, the, people that I talk to. There's. No one person who really knows the end-to-end diagram. Structure so you get a people a bunch, of people in the room and you say well well, how does this work and, it bubbles up the truth it's like a dialectic. Everybody's. Trying, to figure out how how do I send this information and eventually if if you're really disciplined, about getting, this into, a diagram, you're, going to get to the source of the truth and maybe, that means you have to go to subject, matter experts, outside. The room the. Second, key ingredient, besides this, being a dialectic, is the standards, so. We develop standards a couple years ago. Company-wide. For. Analytical, infrastructure, as seen by the business side this is not, necessarily, what the IT teams, need but this is from the business side what. Are the shapes we use the standard shapes the best practices, it's left to right top to bottom and, and. With that we, then take what, we learn from the subject matter experts, put, it into a Visio diagram, that's standardized, knit. Those together and load, it into the IO II. So. Two. Key ways to. Differentiate. A, depiction. Of data flow and this. Is sort of keeping two industry. Terminology but we we deviate a little bit here first of all we simplify, it because incredibly. Complex out there BPMN. And almost other stuff we think about network view you can think about it that's failure, at a single point and you'll. Actually see that in the demo we prefer. The linear view because I like to think of, financial. Services companies like insurance companies as consisting. Of factory, floors, of, tools. Databases. Information. Processing, tools delivery, tools models, delivering. A calculation. To. The business, to. Drive profitability, to sell premiums. Compliance. Etc. So. Now we're. Going to get to the demo and the demonstration, of the product this is this was done, is. There's a video done of a of an actual, ioe. Proof of concept. Tom, Sawyer perspective application. And I. Want to walk through again. In order of complexity, talk, about the fact that it's a process map repository. Perhaps, most important. We're, we're, going to show you how we can take, attributes, and map them to a different dimension and, then we're going to finally talk about taking, attributes, and putting. Them into an, algorithm, so. Here, we are we're in the perspectives. What you're looking at is. 2,000. Object infrastructure. Data. Databases. Models. Information. Process tools data delivery tools and imagine, trying.
To Do what I'm going to show you to do in Visio. Diagrams, on. A SharePoint so. We're gonna select an object object. 504. And it's. It's, popping up there and now. We're going to say what let's isolate we, want to isolate the, upstream. Flows into this object and we want to isolate the down stream. Flows. You. Do that and by. The way our our data flows are captured, as nodes in our neo4j and not that it makes a whole lot of difference you see they're color coded those. Colors actually are informed, by attributes, stored in new 4j we're. Now going to isolate. This. Topology. From. The overall topology, and now, you have the opportunity, to query just this selected, opportunity, that this selected, topology, on. Its own. You'll, notice that, this. Is this is pretty much an end-to-end picture, the databases you see there or the source databases, bring database to, bring data into aog and downstream. You have the, data delivery tools delivering, it to the final, user so, now we're going to switch the. Attributes, that were color coding on we're. Switching now to a. Risk. Attribute. That's found in a number of different authoritative, inventories. Again. These authoritative, inventory. Are managed. And governed already, this is not this. Is not the province of the IO e the io e takes advantage, of the, attributes, that already exist in the company and leverages, of them as, you can see here is a heat map of, the. Ranked, order model, risk scores and, application. Risk scores and database were scores. So. Again, it's you. Can have a multitude, of, attributes, we have in. The case of the proof of concept we we, loaded about I guess, about 50, but we had the option of loading over 300, attributes. So. Now, we're going to focus. On, six. Different tools that are in the far upstream side of this topology now. You'll notice that the. 528. Is a high risk tool, being. Supported, by for. Low, risk tools, so the first question you say well wait a minute are we, is. All the good work and good effort put, into those four tools negated. In 528, I mean it's a high risk tools that means things. Coming out of 528. Are. Questionable. So. That's the first question that asked and again this is something that a.
That. An authoritative image for a silo, table, database, might not be able to show you because you wouldn't be able to see the, topological, connection. By, the same token what about the database is that database, feeding, is, it risky because it's feeding poor data quality, okay. So we're we've. Got a couple questions here and I like, the phrase actionable. Insights, because it fits so well, to. What, the basic, premise. The basic capability, of the ioe is so. We've switched attributes, now now the next attribute we're going to look at in the, in the heat map is, a, criticality. Attribute. And this shows. The. Criticality of those four in the center it's. Fairly critical those need to be brought. Back online within 24, hours after, failure, but, the two on the end best. Efforts, okay. So now we have another question to. Those who are the users the ultimate users, of this how. Critical is this if it's really critical, the. Maybe those green guys should be yellow or maybe even red if, it's not so critical do, we deploy our resources elsewhere, again. This, these are topological. Actionable, insights, that. You can't see. Easily. And not only that but as these attributes, change in the authoritative inventories, those, will change in the ioe so. Try, doing that in Visio it, would be an endless repetition. Of Visio diagrams. Etc. So. Now we're. It's. Escaping, the topology, pulling, back there. We go. So. Now what we want to focus on is the. Fact that this. Let's, suppose I'm I. Own that database and, I'm making a change and, the. Change is, a, simple. Format change I want everybody downstream. To know the, change I'm making so. I can select the object, and. Right-click, and I say send. Notification, to everybody downstream, and we know everybody, downstream, because we're getting that information from the authoritative inventories, and send. It all downstream, and say hey we're, making a change to the format if, anybody's, got a problem let's. Push back. By. The same token what. Can also go to the end user and say to the end user all right now you have the capability, as if, a new, user. Let's say I own a delivery, tool and a new user comes to me says you know I'd like to use your data for. A report, X. So. I would say well let's see if the model owners who are contributing to that, data. Really, believe their model is is. Fit for purpose, so. An, alert, would go out from me is the data delivery tool owner and I'd say is this, use in report X. V. For purpose for your model so again, it creates a community and, this, is something that I've seen. For. Years for decades as I work, with, groups. As either a business analyst, or a project, manager, and having. Them understand, ok this is my world, but. What my world does is feed another world I don't really know who those people are and I don't know exactly what they do this information, this. Allows you the end to end not only not, only awareness, but the accountability. So. That hey, if the data quality I'm getting from my. Community, there in yellow is not. So great maybe I should look a little further upstream maybe the problem is with the data coming into the company. And. Along with community I guess. I reach, a sort of an an. Altruistic, view, of the, use here is that yeah. We talked about propagation. Of risk but how about propagation, of contribution. I like. You know you've been sitting in this in this job for four years and say you know what what, exactly am i doing for the company. This. Is a way to see where, the contribution, of the teams and the individuals, in the company are going once you hook the business metrics, up to. The downstream, side and see okay I've got a factory, floor I've, got five hundred people on different teams working for this and we, produce X amount of premium revenue, for the company, that. Would be your propagation, of contribution.
Okay. So. Now we're, going to go to our network, view and as. I promised this is another way to look at topology, and I. Like to think of this is this is where you can easily see your your. Failure to point risk, obviously. In this topology and, by the way this topology, is. For a single use so, there could be 2030, uses, that are shared among, these. Hundreds, of tools. So. The, question here is all right of the of the nodes that are highly received. Highly concentrated, paths. What's. The risk of those nodes who owns them. Is. Their knowledge sharing going on there's another question for you. Every, company loses knowledge, share but is that knowledge is that knowledge documented. Which, gets back to the whole idea of diagrams, and getting a dialectic, of the truth. Now. We're going to go to using, attributes. As a. Method. To go into different dimensions, so we're leaving the topology, infrastructure, we're, now going to the. Physical geography. These. These happens to be, fictitious. Locations. Of AIG server farms. For. Obvious reasons we can't, reveal the real locations. But. This is what this is where your, tech, risk people would say ok this, is great now I know where my servers, are the connect to my topology, should. I have failure, and the server's. What. Part of my topology, goes down who's affected so, this gives you a vertical, integration on a different dimension, not. Only that but this, is just one of many, dimensions, you could use within a single company, so, think about org. Charts, that's a dimension, that's a topology, think. About the way cash flow runs. In, the company that's a topology, those. Attributes or, those nodes if link to your. Analytical. Infrastructure, then. You could go back and forth between different topologies. Ok. So now we're going to go back. Back. To the original. The. Original view. And. Now. We're going to look at our third and final, use. And this is the one that's our. Third step this is where we went to first, thinking. That we were ready to, to. Get solid support here with. A. Tool. That takes instead of going to a single point and saying. Well what's my upstream and downstream looked like this is now going to a use was. Going to a use of again. I'll call it a factory floor with, two models at the end and some. Information process tools and data bases upstream, and, what. We're doing is we're taking risk, attributes. And we're, using them in an algorithm. To determine, what the overall risk is of, the. Topology, of this of this topology set. So. Again. I mean Shawn Keenan worked on this it was a it's, basically a Gen, Z row kind of thing just an idea, that if, we take the, complexity.
Of, The. Topology set, of the tree and. We. Look at it in comparison, with the risk. Of the, tools, individually. We, can come up with a an. Algorithm, for measuring. The systemic, risk of the tree so. We came up with that but. The beautiful thing here is not only could you do this in a current, state but. When gene built the data model, he added, a level. Of states within. The data model so this is current. State but. You could have a state that is a simulated, state you. Copy this topology put, it into the simulated, state say okay now, suppose we move things around, suppose. We, consolidate. Databases. Reduce. One of those information process, tools then. You can see how it changes your risk. Metric, your. Your. Tree risk. Furthermore. You can have obviously. More than one attribute in your equations, and you can have more than one equation in this. Case we we kept it kind of simple again it's proof of concept but, you could you, could get very fancy, and say well if, a tool is so many hops back I want, to. To. Completely ignore the risk because it just it's gonna be mitigated, away some. Such thing. Attributes. Are. Attributes. Again because, you're dealing with not only attributes, of risk and you're. Dealing with that be, things like who, owns it who the data steward is how. Old is it it's, all it's, up to your imagination as, to what your queries, are going to be but. As, I as I talked to one fairly senior level, exactly. Is. Fantastic. He said it can do a lot of things for this company he, said your problem is not what it can do your, problem is to focus on something it can do that. Saves us money and that's, and that's where knowing, how the company operates, is. Required, so here. This is another tree this one happens to be 58. Models but. Actually has a lower complexity, score and, a lower overall model. Risk score. But. I think, the point on this this. Part of the demonstration is. You. Know another pushback, we get is what you're trying to boil the ocean you're, you're looking at you know the entire infrastructure, of the company yeah. We are that's where the real benefit lies but, it doesn't have to be done all at once, we're. Not talking about putting. The entire infrastructure, in the company all at once we're talking about piece by piece put. A piece in see. If it works get, your benefits put. Another piece in and all the while you're learning, so.
To. Summarize, I, want. To say that you're. First. Of all diagramming. Is this is the heart and soul of the IOE you've got a diagramming, and the standards, that go with it the, attributes, tying the attributes, and tying business, metrics, to the topology the. Collaboration. It's a community, tool, propagation. Of risk but also propagation, of contribution. And. Using. Attributes, to go into different dimensions, and then. Finally using your attributes for optimization, and with. Optimization, this becomes an entry, into artificial, intelligence, there, any questions, Deborah come on oh so. We. We, took a we, took it's off Visio we had a combination so would you have the diagram part and you have the attribute, part the attribute, was simply a an. Extract, from the authoritative, inventories, and that extract, was loaded into Neo I'm. A business guy I'm not a technical guy so I can't go to the level of of. How it was done but, I know on the diagram side it was a matter of taking the. Diagrammatic. Diagrammatic. Topologies, and building. A CSV, file doing a load script. But. But again I mean like neo4j, with, kettle is I. Mean that this is probably the next the next way to go with this because kettle, would be the obvious solution to that problem is, it, does that help. Yes. Sir in the back. Well. Officially and, I don't know if this is an ISO standard but this is level three or possibly, level four I. Mean. Are you familiar with ah. Good. Question yeah I like that question okay, so. We. Started off at the analytical, infrastructure. Level tools. Databases, it's the physical infrastructure, but. What the business, really wanted, was. Hey we've, got we got workflow we got processes. Workflows, work, steps what. We call task sequences. So. We built the IO II and, hats off the gene for making it work in the data model where, we integrate. Workflow. Steps with infrastructure, the. Beautiful, thing is. You. Have on, your workflow, step side you have controls. Maybe there's Sox controls, that's, all on your workflow side not on your infrastructure side the infrastructure, side you, have other concerns. Your. Server farm concerns you have your your. Health and hygiene of your databases, and applications, and, the ioe you can merge those two and present to the business hey here's here's. A combination of your workflow controls, and your. Infrastructure, risk. Well. Yeah so. Yes. Yes. And, what we do in the, IOA is we take each workflow step, and pair it with what infrastructure, makes that workflow step happen and if. We don't if there is none and it could be like you know Joe. Makes a phone call to Jill that's that's. Just a manual step, but. For a lot of this we marry, those two, side. By side. So. Now you can see if you're if you're on if you're on the workflow, step side and you know that's something, we did add to the demo because it wasn't time but we have see, process, flow I could go into the process flow and say what. Supports, this process, what supports this work step and then you can drill all the way down to your server if you want to go, yeah.
If, I if I understand so, if I understand a question you're, like, how can we present different levels so, let, me answer that question in two ways one is it's, it's easily done because, the neo4j, is flexibility, and the way Jean built the data model so that it's, easily done what, we've done so far to, answer your your question more directly when. The business came to us and said oh gee you know you it's great you guys did infrastructure. But we're we're. Really a workflow. Step job. We. Let the infrastructure you know handled by somebody else yeah, so we can show those two separately or or integrated. We. Can't we've Tom. Sawyer was gracious enough to offer us custom, attributes. And. And we'll probably at some point go that way but to answer your question we used, standard. Geometric shapes, which Visio uses as well, yeah. And, the danger was we actually looked at I mean Chris, Tanner offered, us lots I mean a whole world of. Different. Symbols. But we we. Didn't want to confuse the user, with you, know I this is a gimmicky. Fun thing we wanted to stick to the basics, but, who, knows, who knows in the future what we might do. On. The side here if you want to continue, have. More questions come, by. Otherwise. If you can hit the last slide, I think our time is up thank you all very much for, joining us today and see you at happy hour.