First of all, thank you guys for the opportunity um to present um today uh with you this session and share my experience so far in building our customer intelligence assets for E Eisen Bank in the cloud.My name is Anna Peck and um I am currently head of Customer Intelligence Analytics with R Eisen Bank Group and also leading uh uh an IT tribe uh for uh a retail customer engagement where we have uh five cross functional teams developing different kinds of analytical products.

Um So this is short uh shortly about my, this is my first time that I'm presenting at Women Tech Global Conference. Um And I'm very happy to be a part of this uh global community of exceptional women um you know, sharing their experience and also supporting each other. Um So then let me, let me start with the presentation. Uh Please let me know in the chat if you can see the slides, I will now turn my slides into the presentation mode. Uh I will not be able to see myself and um yes, let me just do this just a second. So um I hope that um you can hear me and that everything is um functioning properly. Uh So my topic today is assembling customer 316 cloud to achieve uh customer Centricity. Um a brief outline for today would be uh customer analytical record in the digital age. So what is um customer analytical record? What does it mean? Um How do we then relate um data to the technology? Uh Data is one big concept, right? And the technology is uh sort of an enabler of how we use data and how we monetize data and um how the data at the end brings us value. Um So concept data as a service has been very popular right now, no matter which industry you are, you can relate to this where we want to use data um to basically uh improve our product and improve our customer journeys, improve our services and offerings.

And this really applies across industries. What does it mean to deliver customer analytical record? And again, this is a the general approach to the implementation is really um applicable to, to many industries, right? This is not uh even though I am presenting to you from my focal point of the banking industry, uh definitely the approach can be very easily applied to other industries. So what are the recent developments and trends, especially in the the analytics data analytics world? And when we look to the future where this tech uh may lead us because this is AAA session today that um is mostly about um the cloud technologies and technologies as enablers. Um So where are we today? So with the banking industry, uh there are many uh banking emerging strategies that, that help us uh really influence our up line, upper line and also bottom line, but also this triple line, you know the the people, profits and uh the planet, right? So we try to create a superb customer experience by introducing uh a value uh new value added products as next best action offer for the customers predicting when the customers is unhappy and whether the customer will leave the company and they're trying to maximize the customer lifetime value.

So this is more for the company than for the customer, right? All of these require near real time customer data and usually in an integrated uh uh uh uh layer to enable this customer 360 which is a holistic single customer view of customer profile and preferences. So, I mean in our world, we use data, we create data uh every time, right, by using our smart devices, the variables. For example, this is just one an example, we continually produce multi, multi dimensional data feeds that could reveal our propensity to desire a product to prefer one product over the other, to dislike some service um or a need for a product or service. Um also to perform a specific action, right? That it it might indicate um that we want to purchase something or that we want to go on vacation and uh different kinds of customer insights that we can get from the various data we produce uh uh uh uh for many, many platforms. So what is the customer analytical record? So it's basically uh a concept uh more than uh more than a, more than a tool or a technology um that we use to understand our customers, right?

Um What we see that, that uh uh when we develop the applications, when we also develop data centric products and, and uh uh offerings, we combined uh uh the the the questions, you know, what, who is the right customer, right? So we don't want to even offer our products to the wrong group, right? So this idea of personas and uh uh trying to, to, to create the product vision around who the product is really intended to. And for that also we use data, we use data to personalize the customer offerings, right? So who is the right customer? Um What is the right product for this particular customer? Right? And this can be done not only in the segment uh uh group when you think about customer segments, but, but really the essence of the uh customer analytical record is really to do this at the customer level, right? At the right moment, that's also very important. As I mentioned, all of these customer 360 hubs are in real time. So it it is it is now event driven in, in essence that OK, maybe this product might not be interesting that at some specific time for this customer, but would um you know, there is a time uh generated by some insight when the customer enters specific shop and uh you know, you can target them with specific loyalty um discounts or for example, when um customers withdrawing money on the ATM machine and uh uh he or she is getting a low balance that you can uh write at that minute, right at that screen, offer him an overdraft or something like that.

Pricing, smart pricing decisions are also important. You see this in overall industries trying to collect numerous, numerous data um and model this data to to come up with the right pricing decisions, right? Um So this comes to having a deep customer insight right at this customer level, not at the aggregated level that is stored usually in one place and available to, to be used by various stakeholders across the enterprise in the real time. Um So let me go down further. I I hope you guys hear me because as I said, I'm in the presentation mode. So let me just maybe go closely from this um mode just to see if there are any questions. Um But I guess everything is OK. Right. So that um that you can hear me, OK? Just a second. So I have to then go back, I guess you see me, OK, from the current slide. So basically, uh the the customer profiles we have the, the um as I mentioned from this organization centric setting or in, in your organization centric centric, we had historical transactional data and we usually ask the question, what's the value to company? And as I already touched upon this, uh the customer analytical record nowadays needs to be really persona and customer centric where we can then collect the about preferences, purchases, survey data, uh device usage, sentiment analysis.

And all of these, let's say on a customer level, try to put uh sense uh and, and analyze it together, right? This is a very challenging um process and it's, it's it's many companies um are still not able to have a fully integrated customer analytical level on the customer level where you basically integrate the data and link them uh uh uh uh uh on, on the customer level that that is a very um this is a very challenging but the idea is there um and uh many companies are working towards pooling all of this data um starting usually from the bank, first party data because uh many companies are still not using and utilizing and monetizing the first party data.

Um The ultimate goal would be then to, to integrate the external data, the third party data, right, where you can really relate to a customer on a cookie level. And this is something that it's um pretty challenging uh most, most of the banks that are, you know, no stars in analytics um have are still implementing uh the the platforms where they are able to integrate all of this data together with this um very deep insights, right? At the customer level. So when we talk about data and technology, right? So uh how does one relate to another? So basically, when when you want to create a customer analytical record and you know, this really doesn't matter in which industry you are. Uh then there are certain uh uh technology uh enablers uh that are essential for your implementation. And of course, uh I would start with the cloud computing uh for data analytics cloud is essential. Um It's, it's simple as that. Uh it ensures this um uh uh interop interoperability um uh between all of your applications. It ensures scalability, this concept of uh a scaling um in the cloud where you basically spend how much you need and how much you use. Um And also uh uh you're you're more flexible to introduce more computing power, more storage needs. Um depending on your uh on your current, let's say backlog and in your current uh demand. So cloud is becoming essential uh for data analytics.

We see nowadays, there are amazing uh solutions that come from many uh public cloud providers such as Amazon, such as Azure um that, that really uh integrate the data in very um uh cost effective and uh uh uh with really uh uh great speed and time to market. So besides the cloud of course, modular architecture, when you develop, as you can imagine how many data points are really involved in making a holistic custom analytical record. So you can think about probably some 10,000 data points and these 10,000 data points come from all of the different places and all of the different applications, sometimes legacy mainframe, those are the most difficult to, to extract and to, to integrate. Um So, I mean, in order to approach the data um sourcing for uh uh uh uh what the project like that you would need to think about modular architecture starting from somewhere, right? Planning it um from very simple uh uh uh uh let's say a bottom up approach where you start with streaming and and batch, you need to decouple your data pipelines from the overall infrastructure. Um Because you need to make sure that your compute and storage are uh decoupled.

A good example of this modular architecture and especially in the context of data analytics is the data mesh concept because this data mesh concept allows you to integrate different capabilities. Um And uh uh uh uh is a good concept architectural concept for the integration of the data streams from the various sources. So I highly advise to, to visit this data mash and look into the deeper into this architectural concept for data analytics. One of the enablers is digital identity. The fact that we can compile data about us individuals and also enterprises and organization in the digital form. So digital identity is a big enabler in terms of the, you know, wet identification E signatures, um all kinds of stuff that enables us to really uh enrich the the the data for customer 360 with the digital identity. Um once we have, let's say integrated data at some level, we do have some variety of data sources. We have introduced some um basic customer. Um let's say uh 45 degrees or uh 100 and 80 degrees, then we need to manage this data somehow. We need to make sure that this customer 360 lives, that it's refreshed, that, that the fresh data comes to it and that it stays relevant over time for this. The digital and automated data management is crucial, right?

So I'm only putting here the data ops, which is uh a really kind of uh um uh a new also concept um similar to agile and dev ops for the data analytics. Uh And it's called data ops um to, to prevent the the the the the the the data drift. So data drift is something that is called that, you know, data stays relevant and that it, that it makes sense, right? That, that this analytical record stays accurate. So th this is really uh another uh important technology enabler and there are many uh now cloud solution for data ops that help us maintain our data pipelines in a way that uh the customer 360 stays relevant for us. So these are some of the basic concept of technological concept that would be relevant while we are building the customer 360. In that context, the text tax, um I would start with the data storage and processing. Um usually some sort of a hybrid platform uh where we have the the, you know spark Cassandra non SQL database. You can have also Dynamo DB, which is the AWS fully managed non SQL or for example, Cosmos DB, if you're on an Azure stack, um the point here is that it has it should be in the cloud, it should be really hybrid um that can um serve your needs for analytical but also transactional um aqua uh especially if you're operating in real time, right?

And that there is this real time integration layer with data streaming technologies um that can um successfully integrate into this this mesh, right? Data pipelines, I already mentioned data ops tools here. Um These data ops tools provide real time visibility into your state of the data movement at this time um with alerting uh with monitoring capabilities. Um so that you, you can at any time, check the integrity of your data and your data pipelines, making sure that the data is relevant and that your uh uh product, your data product at the end is uh adequate or uh is correct and is up to date uh with your fresh data uh data integration.

As I mentioned, you need to integrate at some point uh many of these data points together. So they make sense, right, and consolidate them in the essence that this activity is correctly assigned to a specific customer. Um Sometimes um this can be done uh with a customer id, but you need to have some sort of a minimal schema specification or data model that you um use in order to put all of these data points together. My emphasis here is to minimize the schema specifications because if you start um with a complex schemas, then you are building a warehouse. And um I mean there is nothing bad in the in the building a data warehouse. Don't get me wrong, but you want to really start with the basic minimum data model that will allow you this flexibility of integration right at the end when you have your data products, um you want to have some visualization layer uh in order to see OK, what is going on, right?

With the dashboard, right. This is very important. This data visualization layer should be a self-service tool. You can use Power BI I for this, you can use click uh for this. Um There are also many other solutions that I didn't intentionally put anything here, but it's important in the whole process. And while you are building your tech stack that you keep in mind that you would need data visualization and you would need integration with your storage, your Lake Lake House uh with some visualization tool um and uh uh a data product that will be uh supporting this, these dashboards, right?

So in the essence, this is your basic, very basic text stack, you know 101 for customer 360. Um This is all relevant because uh creation of a data as a service is a new term. Um What is data as a service? Um And you would usually find it as a management strategy, but I say it's more than a data management strategy. It's it's trying to monetize the data or build data as a business asset, right? Uh where you would use your data for some targeted business intent for some targeted business value you want to generate, right? The architecture as I mentioned is modular includes diverse technologies. Um Many of them who are that I mentioned here and also um besides the the fact the purpose. So you were building it for some specific intended use or some value, you want to have this modular architecture, what is important and how to deliver data as a service. And this is very important that you have your delivery based on agile delivery principles, you work in fully agile mode and you use this concept of data dev ops, data ops um that are really crucial here for keeping your service live, keeping your service um measurable. And yeah, this is very important for the two questions that I ask here. So before you define the service, you need the drivers and the value why you are doing it.

And also think about the KPIS uh that will continuously help um measure your experience and help you uh um help you with this. Um So how do we do it? Right. So in implementation approach, um in the implementation approach, uh this is a pretty, let's say, straightforward but not as trivial uh in when you want to essentially do it. Um the the best case, how we did it would be that we develop one use case and you start from one single use case and you specify uh what do you need to to make this use case um to implement the use case. What kind of data? Right? So you start with data specification and you um identify your data sourcing needs then depending uh on your organization on where the data resides, um whether you have that data centrally available or you need to get it from somewhere else, you need to organize these data domains, right?

So you, you scan the current environment to understand where the data is. Um And what does it take to get that data? So you know, to gouge really? OK. How, what is my time to data? How much effort do I need to get this data? And plan your data sourcing, then you eventually source the data. So from the external or uh uh uh internal sources, right? Um sourcing the data depends on where the data resides, right? So how trivial it is for us, it was very difficult because we uh r Eisenbach is, this is an international uh regional group that has 13 different markets. And usually you would start with the use case for specific market, which means that most of our innovation would be in our head office, which is in Vienna, getting the data from the local markets. This is already in itself a very challenging task, right? But this depends on your organization and how your data is um uh organized. And um what is your use case, right? Also. So then you develop a data product, uh data product usually requires some engineering. Um uh You start developing your data product for that intended use case for that intended purpose. Um And then you enable some sort of a uh a self service. So this data as a service depending on what the intended use is, right?

It could be a dashboard as simple as that. This is your data product. You just want to have a dashboard or it could be a machine learning algorithm. It could be a machine learning model that this is your data product that um uh tells you about your churn, right? Or tells you about what would be the next best offer or tells you about your pricing um or about your um digital marketing uh targets or uh or per, for example, it could be also micro segmentation where it can segment your customers regarding uh some um you know, special dimensions beyond, you know, the simple segmentation you might have in your organization.

So, but, but what I did here, I basically put the, the the flow uh from how it's uh uh uh how it was usually implemented. Uh I think that this this this implementation approach is very much applicable to, to many industries um creation of a data asset. This is very redundant to the slide that I just explained. Uh one thing that is um message from from from this slide here is this industrialization approach, right? So once you have your use case, you did all the discovery, you did the data, you source the data, you start, you implemented your data product, you engineer the data product, you have your data pipeline, ready your machine learning pipeline. The case you have a machine learning model.

How do you industrialize that? So how do you now? So we did it for one market? But you know this is relevant for all our markets. How do we then uh replicate this? Um And this is very important. Uh I think that this is the essence of the true monetization behind this uh data asset, right? How do we uh then uh put this uh how we replicate this or implement this, deploy this for many of other uh our markets, right? Um So basically, um this is very important to, to, to create a service uh to really think about this last mile. Um So what is that role based consumption? Who will be using that? Um How do you scale your data pipeline? Um And how do you then continue to use this, this this data product in a way that it stays relevant? Um How do you enrich it with new information? Um But to do it in automated way. So I think the from all of our implementation um uh uh let's say uh uh history and our experience industrialization part is always the most difficult uh especially when you have various markets and not one size fits all. So that's also very hard. So regarding the the uh the future and the trends, what are we seeing in the, in the data analytics, as I said? So data in the cloud, right?

Um data uh is in the cloud data will be in the cloud and uh uh more so in the future to come, uh we will see most of our analytical workflows, analytical products in the cloud. Uh Most of our uh let's say when you, when you screen now the the you know the jobs for data analytics, you will see that the cloud knowledge is inevitable, it's essential. Um And this is really important. Um I if you would tell me right now to, to start an analytical project um from scratch with some idea, I will always think about bringing the data to the cloud, right? And this might be difficult for some markets. Um It, it's not all, it's not that um easy in many enterprises also due to the data privacy and stuff. But I think that there is always a way um to try to account for your process. So with these detailed privacy impact assessments, um to really account for your work flows, describe your work flows in a way that it would be compliant with whatever um let's say um uh um uh constraints you might have in your organization, right? Um So big data analytics solutions um solutions. Why? Because um this is an end to end story, right? Um It's not about uh uh um just one data product, it's not about just deployment of a database in the cloud.

It's not, it's about the whole concept of putting all of this data together and, and making sense of it. Um And the workflows when I say workflows here, I mean, this data ops uh this data monitoring uh keeping the data relevant uh applying these agile delivery principles uh uh to our data delivery and to our delivery of data analytical pro uh products. I also put some useful resources here. I maybe this presentation will be available to you. Uh But you can definitely go through these um useful resources. These are some of the, let's say top um to leadership um uh uh um websites uh companies uh that could be relevant for each of these categories that I just mentioned. So, as I said, this is elaboration of, of the, of the uh of these three categories. Um This is the information and the fact that you will be able to find under these resources. Uh So 90% of data analytics solution will be in cloud by 2020 50% of all data worldwide reside in public cloud environments. Um This is uh aws, this is Azure, these are the design, the Alibaba cloud. Then we have also IBM cloud, um 30% of the worldwide data, we need real time processing, right? And I think that this is a even um underestimation 30%.

I think that this is uh in my view, but these are the facts that you will find on the internet. Um This is really growing, right? This this real time part role of the edge computing continues to grow. Um Then when, when you have this much data available on the public A you will also have an evolution of cyber attacks which means that you are hardening your security, hardening your compliance, your data privacy, you need to think about that by design, right? This this is really important then data services and data products. This is what I mean about solutions. So your solution is data as a service or data as a product, it's an actionable data, right? Um Why is this important? Because uh when you have a solution that is actionable, that really uh is an end to end solution, then your adoption is easier, right? You will be easier to um to have your uh different uh uh business units. Use this as a data as a service, use this service, use this product. Um And you will be um you can easily then improve it, you can measure the the success of this product. Um And it will give you some uh some insights whether this is the right fit for the organization and how you can improve it. And uh lastly data culture, right? Uh So the citizen developer, we all heard about it now we have citizen data scientists, right?

So I mean from, from citizen data scientists, we go to citizen developer, what is really important here and what I wanted to say about the data culture is that um most of our applications right now. And when you, I think this is a similar uh experience with, with many of the the companies are very code intensive. Uh You cannot gain some very, you know, simple insights from your data without some development work, right? And this is a challenge for many. So we need to go in the direction where this will be democratized in a way that we have low code platforms. And we will have also people who are not so technical being able to benefit from this. Uh So where the tech leads us, right? And this is a little bit of this flavor about this data mesh concept. I think um we are on time. I want to also leave some uh questions for the uh some time for the questions. Um So data mesh is a really cool concept. I highly um advise you if you are deep into this data analytics um industry, go look it up to works is one of the companies that is really leading in, in this uh to leadership right now in the world. Um The the basic concept is around these four areas where you would have the main distributed architecture, you have this product thinking, you have this ecosystem governance and this self-service infrastructure as a platform, right?

And then examples of data mesh technology enablers are some of the the the ones that, that, that for example, we had experience with. But this is a fully non exhaustive list uh for you to look into um what can help your data match uh um time to market. Yes. And I think we come to to an end of the the presentation. Um Sure you can, you can follow me on all of these social media. Um And um yeah, let me know if you have any questions, right? Well, maybe I can see these questions in the, in the Q and a but they are none right now. So let me know if, if you have any questions just in the chat just simply put it in there. I don't know if anybody is with us and if I went over my time, if there are no questions, I'll just give it a couple of minutes. No. So RB I doesn't require German for recruiting because we are a global uh not global but regional company and we do have um many markets. So we have 13 markets including Czech Republic, um Slovak, Bulgaria, Romania, Ukraine, Belarus, Russia, Bosnia, Croatia, uh Serbia and uh all of these markets, obviously people speak different languages. You're welcome, Michael. You're wel you're welcome. So please um visit me on my linkedin profile.

Connect with me there, Zana Pne, I will accept all of your invitations. Stay connected. Let me know some more if um you are, you are uh interested. Uh No, no, not in Africa. Not, not yet, but I heard very good uh uh stories about Africa as an emerging market and I think it is an economy that could be interested. There were some intentions for us to expand there, but not for now. Yes, sure. Just connect with, with me on link then maybe I can just put my link then uh profile here. Just give me a second. Um Just to make sure you can find me. Let me just copy it in the chat. There you go. Connect with me and thank you so much for your time for listening to me today. Thank you guys. Bye bye.