PegaWorld iNspire 2024: Rabobank’s unique journey to the Cloud

Thank you all for coming. I know I probably will have a few more people dropping in a little later, but, uh, definitely want to appreciate that you were all here on time. Uh, so thank you all for coming. Um, my name is RTIM Whitman. I'm a practice leader for Pega in Cognizant for the northern Europe. And, uh, together with, uh, Tim, we will give the presentation about, uh, Rabobank a unique journey to the cloud. Uh, so, uh, over to you, Tim, to start. Thank you. John.

So my name is Tim Blunck. I'm a product owner in the payments domain at Rabobank and specifically owning Smart Investigate. So the unique part of our journey is about smart investigate to the cloud an existing smart have investigated. It was implemented like 1015 years ago as the first patent application in Rabobank. Now Pega is all over the bank in multiple domains, but smart investment is still there. And we had a challenge because we have like a system connected to many legacy systems. In legacy methods, we had to do a data migration. Well, we're going to explain you a bit more, but um, without further ado, I want to tell you a little bit about Rabobank. If you don't know Rabobank, I will do it little justice because it's only in two slides, but I will do my best.

So Rabobank started, um, as, um, really like 125 years ago, a little bit more with very small cooperations. So it was literally, literally agriculturist, horticulturist coming together and investing and helping each other to modernize or to grow their business. And that basically grew in in the whole 19th century. Yes, or in the 20th century. Um, and that was the origin, the origin. So the corporation started to to come together. Grew bigger. And in the meantime, the whole banking landscape changed. So, um, after the 1950s, it was more common for families to have their own bank account.

So also Rabobank started investing in that, and we started supporting a lot of retail customers. So that is how how the bank grew in the last 125 years. Still, this cooperative effort is, is the base, um, and it's all about like being together. You can do more than, than you are on your own. So some numbers. So Rabobank is one of the three main banks in, in the Netherlands. Um, so we are spread over the Netherlands with a lot of offices. We have um, like more than 8 million of private customers and a little bit under a million of of business customers. And a lot of that is in the food and agri business.

Um, again, um, the cooperative part is, uh, we are not on the stock market, but we have 203 million members. They own member certificates. It's another other way of running a company, I guess. Um, in the blue on the on the right side of the slide is how we are present in, in the world. And that is really in, uh, in lending, in trade and in leasing. So we don't do any private customer stuff there. Um, so today we're going to guide you through our journey to the cloud. First, I will tell you a little bit how our smart investigator stack is set up. Um, then John will take over, um, because John has been helping us at the start of the project to do an assessment.

How are we going to bring smart investigate to the new state or the new new world, so to say, uh, how do we modernize, etc.? Um, then I'm going to tell you a bit more about our migration strategy. We had quite some complexity. Complexity there. So I will explain a bit more how we tackle that and how we got to a successful go live. And then John Talbot will tell a bit more on how we how we do the next steps. So Aztecs. Um, so I said we started as the first Pega application. So at that time it was the first product from Pega.

It is really still one of the products to use by banks for payment investigations. Um, so what we support is a lot of standard payment investigation functionality. So on the left side you got the investigations for payments, um, investigations for Sepa, which is uh euro payments in the eurozone. Um, so there's a lot of standard processes that we support there. And basically what we offer there is a lot of standard ASI functionality. We offer basically the user a palette of options which they can use. We guide some, but it's not fully guided. On the right side we have two processes. One is reconciliation.

So we do that on nostro accounts. So interbank accounts but also on internal accounts. And that together with the savings process which is really a tax regulated savings product and that is fully workflow based. So we really guide the users through that. So there's a big difference between the two. There's also a big opportunity for us to to improve there. And then we do some other payment related processes. And there are some more coming soon. Those are not specifically that we need the ACI platform, but we got like the connections, we got the users, we got the access control.

So we are just integrating them. Um, so that is a bit about our, our setup. Um, then the uh, the features that we offer, a big important one for ACI is, of course, the connections with Swift, um, and all the messaging that you do there. Um, we do a lot of messaging to, for example, the customer secure inbox in that online environment, But also outside of the company to other parties. Um, and I think in the center of this is this online search portal. So basically what we do, we there is a case initiated in Aceh. And what we do then is we start enriching that with data out of our payments domain. So we have a lot of connections with all the, um, the applications around us. And we offer this to our users to start their investigation or continue their investigation.

And when they they come to a solution, then of course, they want to do transactions or they want to do messaging over Swift. So that is what we cater in this product. Um, what we try to do is that so as steep as possible. But that is a bit for later. Um, John. Yeah. The assessment part. Thank you. Yeah.

Um, as I, as Tim already said, right. Has been around for, for a long time and we, um, we are supporting Rabobank also for a long time already on, uh, on this domain. Uh, so we have teams, uh, of Pega engineers working together with, with Rabobank, uh, in maintaining, developing the whole environment, um, as I also smart investigate also I think is is one of these applications that a lot of, uh, banks are using, but I think there might be a common history in terms of, uh, um, how to build the roadmap, how to prepare for the future. Uh, it's been around for a long time. Uh, we know that we've been upgrading also multiple applications through the, through the years. So at some point in 2020. Uh, the question also from Rabobank was, how are we going to prepare for the future? And together with, uh, with Pega, uh, Cognizant and Pega, uh, did a number of, of assessments to, uh, to look at, uh, the different aspects of, uh, how could you move to the, to the future? So, uh, this is the overview of what we did, uh, across 2020 and into N2.

21. Sounds like a long time ago by now. Uh, it was all, uh. Well, I wanted to say bc, but no, it was not. Not actually BC right before Covid, but during Covid. Uh, and um, on the one hand, we looked at the, at the, the processes. So the operational, uh, work, uh, walkthrough. I'll talk about that in a, in a moment. Uh, and then we matched, uh, the gaps with the, the future needs with the roadmap of, uh, smart investigate that at some point was also called, um, payments exceptions I think at some point.

And then uh, on the more technical side, we looked at, uh, at the Cloud feasibility and what it would mean in terms of, uh, running the teams, the delivery, uh, consequences for for running everything from the cloud. So after we did, uh, the whole set of, uh, assessments and I'll come back to the operational walkthrough, uh, the choice was made to to first go to the cloud. So that's where the journey to the cloud actually started. I love it how animation always makes you look a bit younger. Um, so it was a hard choice. Yes. We have so much to improve in Pega. Um, but we also had so much time that we spent on our lifecycle management. Um, so the choice to go to Pega Cloud first, knowing that it would take us quite some time to do, um, well, it might be hard is not really getting excited from that.

You do a lot of stuff, but in the end the user just logs in and doesn't see a difference. Um, but anyway, that's where we started. So why did we chose for cloud? Or more importantly, why did we choose for Pega Cloud? So as I said, this lifecycle management, um, that takes us a lot of time. And we in the payments domain are like one Pega entity. So we we have a lot of Pega developers, which are not necessarily ops guys. So everything you do, you need to either somebody to help you or you need people to really stretch and you need people with a very big t shirt size so that they can also cover that kind of work. Um, then a good thing is that, in in big part of the bunk, a lot of teams are already on Pega on Pega Cloud.

So there's a lot of knowledge. And that makes it also possible that we can share components. Again, if you're one team, it really helps if you don't have to do everything yourself. Um, and uh, also good if you go for the Pega Cloud could also support our data migration. Again, our team page developers, everybody knows a bit about data migration, but it's a real thing. It's a it's a big thing. So really good that we could do that. So that's how we we got to this, uh, partnership. Um, okay.

So we made the decision to go live and I think we go a bit more with, with our feet in the mud. Yes. This is not like the fancy stuff. We we we go for GenAI. This is like, really getting the basics ready. We are. We're in a pretty good shape. We were on eight C 8.4, so there was not so much to do there. We were also not heavily customized, so there was also good.

Um, but I came in as Bo and one of my task was to ensure that we go to the Pega Cloud. Um, so my my idea of a of a, a big project like this is like, you go first. Yes. You put something live, you have some users on it, you have you have some fixes and then you slowly build up. But we, we saw that we had some issues there. Um, for example, 90% of our processes need historical data. So either your database is in on prem or it's in cloud. Um, splitting it was really hard. Pega strongly advised us not to do that.

Um, then we also had 34 interfaces that we needed to move. Um, some of these interfaces are fairly simple. It's an API. It's, um, a connection that you do for data enrichment. Um, you can easily set another, uh, subscription up and you just put that live in both environments. Yes. Um, but some of that I really like in the payment factory where, um, that for example, a payment engine really needs to do work to get you in a parallel situation. Um, and then you have so many and you have, um, a payment factory where so many things are going on all the time. You're fighting for priority.

Yes. Um, so we were really in a situation that we had to look for. How do we, um, get autonomous in that respect? How can we really organize things ourselves? So that is basically our our basics. When we started for this plan, there's something else. We are in the payments domain. So we have high value transactions going through our application. Um, we have users with a heavily controlled user access.

Uh, of course we don't want people to be able to do too much, uh, for principle six AI principle, all kinds of connections that we have highly secured. So not on connection level only, but also on the message. Um, so we have to use ASM Luis all kinds of structures and then different for multiple applications because we basically connect to anywhere. So we had to come up with a strategy. How? How do we go live? We were sort of forced to do that in a big bang, to do the data migration and switch everything. And how do we de-risk that? Um, because of our AIC rating, it's like two, two, three, two.

So we got like an availability of two, which is extended office hours, um, integrated integrity of three. That's why we do design, messaging and availability. Um, sorry, the confidentiality of two. Um, so we so because of that, um, we had all these dependencies, but we have to, to go for the big bang. Yes. Um, so we start thinking about what if we take out a few of the interfaces that we can really own ourselves. So we picked out the file transfers and said, like, what if we build a funnel, we let those suppliers or those applications that we connect with, we basically don't even have to tell them that we change. They just send us to their files, to our on prem and we just route it to cloud. So in that respect, we could directly already take eight off fully our own control.

We can switch when we want. We can can really control that. Um, that funnel we also use for other purposes later. And I will tell you a little bit more about that. Then we had like old integrations. So it was integration with uh, SSA sa system from Swift, which is fully on MQ. So try to connect with MQ to cloud. That's not going to work. Yes.

Um, asking them to to provide an API was not going to happen. So we had to come up with solutions with bridges, um, layer that that is exposing APIs to us. Also not very future proof, but first, first things first, I guess. Um, and what we did is we already created that in on prem and then started running that, um, and had really like volume over it. So we were sure all the child diseases were out with it and we could just simply switch at the moment of life. So that is what we did with some of the interfaces, um, in on prem to, to de-risk. Um, and then really on the data migration. So I think the connection between on prem and cloud, um, was really a big thing. What you want to do with the data migration, you want to test that on performance and on integrity, and you want to do that with production data.

Of course, you can do a test with with your first test with test data. But if you really want to feel the pain and find the odd ones out and and want to be sure that people can still work with your migrated data, then you have to do that in production. So we did that. We had a production copy of our database in, in, on prem, and we connected to our cloud production. And we just started doing multiple data migration. And that is where where Pega came in and they supported us big time to um, to make that happen. Um, we had to do quite some tweaking to get the performance because we were only without availability. We have a weekend to get that done. Um, and in the end that worked.

I think we did 3 or 4 full migration runs there. So we worked on de-risking and then we really go to the execution. There will be some duplicate here because this is really the work and some new stuff. So I will touch on the highlights there. Um, so the assessments were done initially in the assessments that I was talking about. What do we need to do to the cloud? We had to do some modernization. We chose to go to 8.8. Um, that also gave us some extra work.

Um, so the compliance report tanked a bit, uh, when we did that. So we had to do quite some upgrading there. Um, we knew we were cloud compatible, of course. And, um, well, we had some modernization work to do. Um, the upgrading of the interfaces. I told you a bit about the de-risking that we, that we did and the unsupported integrations. Um, but we also managed to do that. The ones that we didn't already run in production in on prem, we just started doing that in cloud and started already testing production connections. And you can fairly safe do that as long as you don't send anything out from your, um, from your cloud environment and specifically with, with the interfaces, for example, with your Siebel, where you do data enrichment are fairly good to test before go live.

Yes. That is also where you bring your users in. You let them do some of these checks for you. Also the checks on the migrated data. So one other thing that took us a bit. We are Scrum and agile scrum team. Yes. So we do small implementations. Uh, and it's all pretty easy to control.

But if you start working a year going to the cloud, you start modernizing. Um, but at the same time you're running production and you do like production fixes and production upgrades and then then your code integrity is is really important. Yeah. So our team really had to step up because we had to be much more in control. Everything that is just not fully ironed out, that will break in a process like this. So that was really interesting to see how the team measured in that respect and many other, um, occasions also. Uh, but also we had to do tracking and retrofit, if you know how these rule based setup is, you've got like this whole rule history and basically in the cloud, you want to sort of retrofit it to to one new base baseline basically. Um, so we use multiple environments to go in stages to this retrofitted code. Um, and that worked out pretty well.

Uh, let me see. So on the data migration, uh, one thing maybe. So we did a data migration and we had real users going into production and really test can I still do my work with migrated data? But we also did that as a shadow run with some of our critical processes. So we have one monthly certification process, for example, where every internal account has to be verified. Is that balance okay? And the owners need to do all kinds of approval processes. It's really a compliance process. Um, so what we did that runs monthly, we took the file after it ran in on prem, and we just moved it to cloud production and ran the whole process and let real users, a selection of real users do their work in there.

And what you find is because, um, this is typically a process that went through a big part of the bank. So the whole authorization logic, the just being able to see what you need to do and approve, uh, those checks we find found quite some, some issues there and that really de-risked our go live because my biggest worry was like, okay, we go live. If you go live, you do a big thing, you will have issues. But what happens if you just have too many issues for your team to fix? Yes, it just takes too long. We have daily processes that are really high value. If they don't go for out the same day. For example, we can't connect to Swift for whatever reason, we are directly confronted with fines and that kind of stuff. It's really impactful.

Um, so the Shadowrun was was really a big thing for us. Um, and it worked because we are live and it just now it was a bit like tense for me because we went live on the 21st of May, so only a few weeks before, um, and we were already invited for this, uh, for this, um, breakout. So it was a bit like, okay, what if what if it doesn't work and we have to postpone. But but it worked out and I must say, um, the whole de-risking thing, in the end, everything fell together. We had planned for 3 to 4 week aftercare time. Um, and I think after day three, we decided that a two times a day check in with our users was not needed anymore. We had like 20 or 30 issues, all qualified as low priority. So it was really if I could have signed this for this like a year ago, I would do and I would probably even pay it a bit for it. Um, but okay, you're in Cloud.

And I said this was a painful point. A lot of time, a lot of work, and our users are basically looking at the same thing. So now is is basically the baseline is set for us to do all these nice things that you can hear all day about here in the, in the conference, all this water, water filling, filling things. I guess what we what we already have planned is the second part of Swift ISO. So we did a big migration, uh, for um, the uh, target, uh, payments. Uh, that's the European scheme, um, for Swift uh, took us about six months to do. Was also really hard to do. Of course, with this whole, uh, cloud migration plan, we, we hope to to have been in the cloud earlier, but Swift ISO came in between. So we will finish that.

Uh, we got that planned starting in two sprints, I think. And then the really fun thing started, and that is where I give, the stage again to Arjun. So yeah, this. This first goes back to, uh, the operational walkthrough. But actually we plan to have it in this sequence because basically this is where the. Yeah, the, the business benefits the, the what can you do on top of going to Cloud really started. So bear in mind this was um, three years ago, uh, and uh, maybe some of the advice might have been slightly different, but, um, we did the, uh, the walkthrough, um, using the, uh, operational walkthrough approach that, uh, Pega also performs quite often. We did it together with Pega, actually. And for people who don't know how that works, you really sit together with the operations for a while, understand what happens, uh, during the day, sit with customer service, and then, uh, at the end, you start to note certain things and you start to raise questions.

You have the the conversations with the SMEs, the subject matter experts, to understand why things are the way they are, etc. and then you start to identify where could it be improved, right? What? What are the possible challenges, opportunities and how could you perhaps suggest improvements against that? So all of that was documented in terms of process understanding challenges and improvements as as one deliverable from this walkthrough approach. Um, these are the main findings. I'll not try to walk through all, uh, six of them, but basically findings are connected to recommendations. So I'll, uh, go mostly through the recommendation side. And roughly there are three main things popping out.

The first one is that, uh, better integration between CRM portal and the investigations tool. So to really improve the journey for the customer service agents to really understand where the request is and together with the people who do the the the further investigation to have a good, uh, back and forth and a good customer journey. Then there were a number of out of the box features from, uh, smart investigators that were not really used, things like, uh, getnext work. Um, but also some of the, the templates that, uh, are there already to have, uh, standard correspondences, uh, um, implemented and then, uh, some recommendations on the, on the different, uh, formats and the different messages used in particular. Uh, Swift was used for certain, um, uh, payments that could be done much cheaper through Sapa. So some of those recommendations were also in there. And this is, this is, let's say, all the kind of feature improvements. And then actually, of course, there were also a number of process improvements, uh, that, uh, that were also in there. And uh, we'll have a look at those as well.

Yeah. So these are maybe not the slides I'm most proud of because it shows how much low hanging fruit we have. Yes. Um, so we had this user interviews and we came up with, with one of the two like big projects that we have process client and query. So if you see the blocks it's not really about the details, it's more about the colors. If you see the red blocks, it's manual. Um, orange is guided and green is automated. So just by spending some interviews and start looking at this, we came up with this. And that's not like a month of analysis.

This is just like low hanging stuff. I think one of the issues that we had, why we didn't already do this, first, we of course we wanted to go to cloud first, but we are not in Infinity. We are just like in an old way of coding. It's another issue that we have. We need to upgrade there. It's really hard for to hire or find a developer that gets excited or even understand how our code looks. Yes, what is calling this agent is suddenly very difficult if you're not used to that. Um, so another process, basically the same thing. There is really a lot of opportunity.

So you can understand the last two years when I talk to users, I just try to give them everything just a little bit to just to keep their life alive. Sort of. But it's not a it's it was not a fun job because I know we are just neglecting them big time. Um, so I think after the assessments, what we did and at that time there was no Infinity in, in Pega as I at all we saw in 8.8, uh, one of the, the the blocks there is having Infinity in there. So we now see it also showing up. And I think in the short future we see a lot of uplift there. But we did a POC to check out like is it possible. And we were able to do that. So we decided, okay, this is also what we want, because this is what Pega developers coming from school really love.

Yes, and this also makes it possible for you to use App Studio. App Studio for us is just like I have my BA's. They went on the training, they are certified but yeah, there's nothing for them to do there. Yeah. So um, so that is really going to be exciting because there's a lot of wind. But I guess even then, um, Constellation Infinity is something we really want to do, but it's not a goal as such. It's just going to be slowly going there because we need to just do something on our user journeys. And one of the things that we really need to optimize on is the operational excellence. So really the spreading of the load, assigning cases to to users on, on their skills, on their, uh, role level, etc.

really basic stuff. But um, I know from also other C users that it often still happens halfway in a spreadsheet or even on paper, because it's just hard to get all this team organized and at the same level. We are lucky. Our payment investigation teams are looking to combine and they're really working and working out who is able to do which task and that of course, we are going to directly get from them. And in that respect, we can really help them in that change. Um, so you saw it in the in the two examples. Yes. There is a lot of STP possible. So we will start optimizing those flows very quickly.

Um, um, and we, we are going to start our employee. So we have employees literally I heard somebody saying like people going home with RSI because they have to click and they have to scroll so many times on their mouse. It's just, uh, well, I think when going from empty to. Mm. So the new Swift ISO that helped a lot. It's much easier to read and stuff and easier to, uh, to oversee. But, um, yeah, we really have a job to do there. Um, and then product awareness. So I think that also counts for me.

I've been focusing solely on, like getting to the Pega Cloud. That was my job. That's getting priority everywhere. Getting all these integrations somehow aligned, getting people together. Um, so I just didn't want to think about GenAI or whatever. The good thing is, we are on 23.1. We are using the layer of distribution and other domain in the bank. So we are on 23.1 and hopefully we will be on 24 later this year. Um, and we are very closely working now with Pega together to align our priorities to also to the Pega roadmap so that we do it in an optimal way.

It would be fairly stupid to start rebuilding something in Infinity that in in half a year is coming out of BIX, so we really want to do that closely. The good thing is, we have so much low hanging fruit that it's not too hard to just align that and iron that out without too much waiting time. Um, but also for the users, because the users, what you do is you start talking with users and some of them are working like for ten years in this these applications so they know it better than us. Yes. So we also use these users for the user testing. It was really a big part of our success. Um, for, for the go live. Um, but we need to like sort of like, um, intrigue them. What is possible?

More so we are trying to start up with demos and show like, okay, this is also how it could look like for them. It's really hard to to realize how they how they can improve. They would really go in what they have now and improve on that. So, um, we already started our first user meetings last week. Um, so we got a whole plan rolled out to come up with a top ten, 15, uh, of their biggest wishes and then start working with Pega how to, to realize that, um, so this is really the exciting work, I guess. Um, so how that will fit in the roadmap? Um, we'll, uh, we'll tell a bit. Yeah. Being here for, uh, for two days already and listening to, uh, to all the, the new features, uh, made me think very much about, The recommendation was already from three years ago.

The plans that you kept alive with the business. It's coming, it's coming. Uh, we're also, uh, already from a couple of years ago. But if you look right now, uh, what can be added and especially, I think, for, uh. Yeah. Smart investigate domain. It's quite a lot. It's quite a lot and very applicable. So on the short term for the 2024, uh, roadmap, there is, on the one hand, uh, a number of uh, industry compliancy improvements and in particular the latest Swift standards will be implemented.

Uh, but on the right hand side, there was already an announcement of of smart investigate next gen, uh, which basically starts to include all the all the features of uh, um, uh, the whole constellation, uh, front end, the the way that you can use all the GenAI tools as well. And even more out of the box process automations. So, um, you can see that there is a lot to gain. And in particular, while I was, uh, listening to the closing from, uh, Don Schuerman. Um, you can you can think about how, uh, payments, uh, are actually very suited to have everything applied. You can think of the, of the Knowledge Buddy to explain what is happening. You can think of the Coach, the GenAI Coach to also help do the investigations and give advice. You can think of process AI to uh, also start to, uh, give some predictions about can this be, uh, closed or is there some suspicion. So I think a lot of the, uh, the tools that were front and center in, um, in the two days past, uh, can all be applied.

And if you zoom out a little bit further, then there is, of course, other technology for identification. Uh, A AI also to identify the the signals that are basically creating these smart investigate follow ups. Um, and of course, uh, the uh, uh, the Infinity front end or the API to start with, uh, in the case of, uh, of, of Rabobank and finally, the point number four here, uh, I think, uh, there are a lot of places where, uh, fusion delivery maybe not, uh, called the same thing in, in North America, but basically adding to Constellation the reusability of components and making sure that you optimally use App Studio versus dev Studio, and that you make the whole implementation a little bit easier to maintain and to understand also that I think, uh, will give a lot of, um, um, you know, mileage, uh, to, to provide all these, uh, business updates. So indeed, uh, we were not the first one we saw to talk about infinity and beyond. Buzz Lightyear was already on the main stage yesterday, I think. But in general, uh, we, uh, we believe that this is a really good, uh, future for sci. And, uh, lots, lots of things to to do, uh, to to take it forward. So that leaves ten minutes for questions. Um.

There's a microphone. There are two mikes there. And. Yeah, we can also. Yeah, if you can help. Megan. Thanks. Hi. Uh, did you have any concerns about data privacy or data ownership?

Um, not really, but that is because in Rabobank, there are a lot of teams on Pega Cloud already. So there's a lot of contracting and a lot of pre work done there. And there are a few requirements, uh, for example, where uh, where the cloud is hosted. So we had to I think we hosted in Dublin or Frankfurt. I don't know what the end result of the discussion was, but that is some regulations there, or some reasons why you would want to keep it in your part of the world. Maybe, uh, but that is uh. So no, there was not really concerns there. Um, for us. The data was moved to the cloud entirely.

Yeah. Oh, yeah. Yeah. So the whole data. So you mean so you meant security when we moved it or. Uh, well, also, because I was thinking if you kept the data to yourselves on premise and the cloud was just consuming the data from your. Yeah, yeah. So there's basically two things. We were on an Oracle database, so that complicated it a bit because in the cloud it's in Pega Cloud, it's Postgres.

It's sort of similar, but you cannot really mix that. Uh, I think you can also create a solution for that. But then it would also mean we would still have the, um, we need somebody to host that for us. Yeah. So so we have again that dependency, So for us, it was easier to put it in the cloud. And since it was legally completely covered for Rabobank, there was not a really concern for us. I think the concern getting it completely over and getting a full secure connection to just let Pega get in your database and just pull the data over and then do it, convert it in the correct way. That is where we spend a lot of time on, but that is a little bit different than what you're describing. Okay.

Yeah. Thank you. Any other question? Yeah. Over here. Or maybe you saw someone else. Yeah. Sorry. Thank you for the talk.

Um, so I have a question regarding what you mentioned about, uh, the Big bang approach for migrating, and, uh, you refer to having a high risk of perhaps encountering multiple issues. So I guess, uh, what was your mitigation strategy there in case you face those issues? Have you had a rollback strategy plan or something like that? Yeah. So there was there was no rollback that made it really interesting because as soon as somebody works in in Cloud, then you can't sink back anymore. So what we did. So there were a few things we did. So we used a lot of user user testing a little bit in production, but a lot in the test environment of course. Um, and um, that, that is really like, well, maybe the test process.

I have to start a little bit earlier. What we did. So we started testing in, in our test environment in Cloud. So we did it ourselves. So we had first our tester doing the whole thing. Then we did a test with the whole team, took two days and just tried to break everything. There were findings from that. So we did a ton. We took like a few days or a week depending on how many issues we had fixing these all.

And then we hand it over to business users, and they started doing that and we had a few key departments. They went through all of it, and they did all their nitty gritty exceptions and all the stuff that we don't even really know. They do. Yes, they know sometimes the application better than us. So and they kept on testing that. Then we did fix it again and we repeated that three times. And then we came in a situation we didn't really find anything anymore. Then we also under migrated data. We did some user testing.

Um, that was a bit more limited. And also with the Shadowlands. And then in the weekend we had this whole runbook set out where we closed off, um, as in on prem, and then started slowly switching some switches. And then we had users coming in and they would do penny tests on the most critical connections. So we did a penny test with Swift and we waited for the message back. And so and slowly we built it up. And then we had this go no go. And that was really the point of no return. But we were able to, of course, very easily go back to on prem if we decided that in the weekend, that would be a nightmare for me.

I think I would just go to to the Bahamas for a few weeks in. But it was possible because then basically you left your on prem as it is, you had to switch some, some, uh, some interfaces back. But um, we also had quite some, um, other applications that had to do an implementation for us. So they often did that on Friday night, and we agreed that we would call them on Sunday night and then say like, okay, I call you standby. They need to know how to revert. Um, luckily that was all not needed. Um, but I think this all together came into a situation because you had all your interfaces running, you had real users on it. All your access control was checked. So that was the best we could do.

And and it worked out. We just had almost no big issues. Right. So it sounds it sounds like your mom was having flashbacks of everything that that happened. Right. So yes. Yeah yeah yeah yeah. But wants you to relive it too deeply. Yeah.

It was really a travel, but, um. Yeah. I saw one question here and there. Uh, still. Um, I have a question on the data migration. Um, so how big was your database and how long did it take to actually migrate? Yeah, it was actually not that big. It was, uh, 775GB or something. Um, but what you find out, if you start pulling that over, that some of the tables are really big here.

Attachment table, for example, is really one in a big size. Um, and that turned out to be sort of a bottleneck. So but that was also one that was easy to take apart. So what we did in the week before the Go Live weekend, uh Pega already moved that over. I think they took like 40 hours for that. Wow. Um, and that is pulling it over, converting it and then putting it there. Um, and then the weekend itself, they took 20 hours for the rest and then do some catch up. And they had also to do an upgrade there.

And Pega has sort of a, I go a bit in more detail than I fully know, so don't be too critical. But they do basically a clean install when they do the whole, um, data migration, they migrate it and they do it on a completely clean slate. We already had a live production, but it just completely took it down and set it up again. That is, I think, the best way they decided to do it. So we trust them that they do the best and it works. So that was good. But in the end, it took the 20 hours in our migration week and that is in our migration weekend. So that was quite okay. Um, okay.

So then you also had some delta migrations after. Is it because you did like big before? Yeah, one week before. We started with the the attachment table. And that is an easy one. Yes. That is sort of stable. And you can easily catch up with the ones you missed in that week before. Uh, and then we just compiled that together with the, uh, with the weekend itself, with the rest of the tables.

Yeah. And that was just to get into this 20 hour, I just I gave a wish. I want to do it in 20 hours, because then on the Saturday I can do all the switches. I can do the testing if we can. If we have problems, we can do a lot of fixes until Sunday, like 6:00, and then we have to make the decision, okay, we don't go. That scenario never happened, but I wanted to buy this time. Yes. We also closed our business a little bit earlier on Friday. It was also a whole battle for them to stop at three instead of 5 or 6, because they have to really plan that.

And well, anyway. I'm a oh, okay, so maybe you answered my next question. So your cutover was basically that Friday. You stop your operations, do your final migration, and then you do all of your migration and all of that and then okay. All right. And on the interface. Right. You said some unsupported integrations. Yeah.

Can you explain a little bit on that. Yeah. So we so this this whole payment domain is continuously improving. Yes. But some of these applications are really old, so they work with MQ or they work with Connectdirect or so these things are you need to either upgrade to a new version if you're the one that you are connecting to, can support that. If not, then you have to to build something in between. So we found a team called CMA. Um, and they basically they consume that MQ, uh, in this case MQ and they, uh, they, they, they publish it to us as an API. And then you just connect as an API and then it's suddenly easy.

Yes. Um, so we had a few of those integrations. We still have one more questions and we have one minute here. So maybe we can move to the last question. Thank you. Yeah, yeah I think the lady over there was, uh, was first. Hi, gentlemen. Thank you so much for this session. Uh, my name is Radhika and I have a question for you.

I have two questions. Sorry. Um, the first one is when you decided to go from your prem. Be it cloud or, uh, the prem to the Pega Cloud. You must have gone through some POC. And I understand that being a European bank, you have the all the legislations, all the restrictions, all the policies of Europe. But apart from that, being a financial institution yourself, you also must have your own policies regarding security, your own, um, things around security. So what were your, uh, boxes to tick when you were doing the POC while moving with respect to security, especially when you are going to the Pega Cloud? Yeah, yeah.

So I think I'm lucky there or we were lucky there. So there was an assessment done by, uh, together with, with Pega and with Cognizant on what is it technically possible. Um, and but we already had teams in distribution on Pega Cloud, so they basically did that whole thing. Of course, there is a whole process that you had to have to go through with contract management, but also with all kinds of risk products that you have to deliver, which is basically standard when you do either a cloud or an on prem. You start with a new supplier or a new entity with a new supplier. Um, so yeah, that was not really a big thing for us. Okay. Thanks. And the second question is you mentioned that you have you have the agents who are using or the business users who are using your applications for over a decade or more.

And then you had to change their mindset to use this new tool. How tough was it and what was your approach of introducing that tool to them? Yeah. So I think we just started that journey. As I said last week, I had my my first meeting. Well, I think the good thing, which is really a bad thing because I was not able to give them a lot like last two years. Yes. Continuously say no. Yeah, we have to do Swift so we have to do Cloud.

So when we just started like, um, next week or next step, if we just explain a little bit about that, you see them like, wow, this is so nice. So they really are inspired. But we just starting this journey. So we will have multiple sessions in this month. Um, just like showing them what is possible and then letting them pick what they want us to focus on first. Um, it, it makes me feel a bit bad because they are so grateful that we want to do something for them because we treated them so bad. It's just like, not a. But anyway, it's an opportunity and we really want to work their trust in this way to make their life better. So yeah.

Anderson, thank you so much and good luck for your program. Thanks. Uh, we're over time. So I think, uh, probably need to to to wrap up. We can do some questions afterwards. There's no problem. Yeah, yeah. Okay. Thank you very much.

Thank you very much. Thank you.