PegaWorld 2025: Banking on Innovation: How Lloyds Banking Group Transformed Fraud Servicing with Award-Winning Digital Self-Service

PegaWorld 2025: Banking Innovation – How Lloyds Banking Group Transformed Fraud

Welcome everyone. Hi. I'm Paul Conlin. I work in the go to market strategy team at Pega. And good news, you're on the home stretch.

The last session after a packed two and a half days. But you've left one of the best ones to the end. Today we're going to hear from the team about their, um, how you use Pega to deliver an award winning self-service application. With that, I'll hand across to the team. Okay.

Thank you. So just to introduce ourselves. So I'm Elliot. I joined the bank back in 2020. Sorry.

2023. Um, I was actually 2013. What have I got? Got that wrong. Working with Pega for eight years and, um, I'm sort of focused on retail customer products within the bank. Yeah.

Hello, my name is Simon Checketts. I am the head of product for, uh, the self-serve journeys for Fraud and Dispute on our digital app. Um, I've been in the company for far too long. It's embarrassing to say I've worked for 30 years in Lloyds. Um, I did have her when I started.

Um, clearly that isn't the case anymore. Probably work with Pega for around six years now since we migrated over to the to the cloud. Okay, so I'm Samer Nazer and I had the engineering teams in Pega consumer in Lloyds Banking Group. Um, I've been in the bank for approximately 25 years and it's all good. So, um, most of the years I spent is all around BPM workflow.

So I've been working with Pega since 2012. So it's approximately 12, 13 years and I have a good handle of the Pega technology in the bank. Okay. Thank you so much. So I'm conscious not everyone here will know who Lloyds Banking Group is or what we do.

So a bit of a brief history. So we are a, um, a financial services group. We're based in the UK. Um, we have about 330 years of history. Um, so we started around about 75 years before the Declaration of Independence was signed, um, and actually in English style.

We actually started one of our brands, Halifax, in a pub which is still there today. So if you want to go and get a pint and actually do karaoke there as well now, so a little bit different, um, you can. Still do that. I've not been. Yeah.

There you go. Um, we are the UK's largest mortgage lender, and we're also, um, I guess on topic today, the UK's largest debit card issuer. Um, we have 28 million customers. And of those, 23 million of those are digital or digitally active customers, um, 46% of our legacy IT Citi estate has now been migrated to Cloud. Um, and I guess for a bank of our size and our scale, that's, that's already kind of getting there in terms of our Pega Cloud or our Cloud, um, progress in terms of Pega, we're about 90% plus so far, and we're hoping in our area to be get the last couple of applications in by the end of this year.

We fully Cloud off prem, um, we have 16 brands and they're supported, um, by quite a number of colleagues across the bank. So you can see there our 16 brands there, and there's a little Pega logo that you can see there, which um, denotes that P&G is involved in that journey or supporting those brands. Sorry. Um, our brands span many, many things across the bank. So we're very, very UK focused.

But we've got our three core retail brands, which is Lloyds Bank, Halifax Bank and Bank of Scotland. Um, we also have um, MBNA, which focuses more on credit cards and more recently savings products as well. Um, we've also got other brands like AMC. That's the agricultural mortgage company, and they support UK farmers and doing things like securing land. We've got Schroders and Embarq.

They're more into sort of investment and wealth products. And we also do um, we've got a couple of brands that focus on car leasing and financing. So quite broad, but very much focused on the UK economy. Um, we have a financial relationship with over half of the UK population, and we're very focused on, as I said, on the UK economy. Um, the single objective across all of our brands is helping Britain prosper.

Q thanks, Elliott. Um, I guess just, just, just in terms of what we're going to talk about today. So we'll sort of shortly talk to you about our fraud, self-serve, uh, journeys that we delivered initially, probably about 4 or 5 years ago, and more recently, the fraud Journey. Um, but before we do that, I thought we'd just do a bit of a double click around the organization and how, I guess we sort of set ourselves up to, to deliver technology. So, um, we do have a Platform structure in Lloyd's.

There's about 20 platforms at the moment. Um, each platform's got around about 5 to 6 labs each. So quite a big operation to manage all the, all the tech. Um, our platform, as you can see, is the economic crime prevention platform where the good guys in the bank, we're a force for good, um, where we can we're trying to sort of stop fraud. Um, and our vision, I guess, is sort of threefold, really.

So first of all, um, I guess become more effective by giving colleagues and I guess, our systems, more information to help them make the right decisions for customers. Um, I guess secondly, it's about developing world class decisioning tools to stop fraud. Um, and then finally, just how do we use tech to, I guess, create innovative and simple solutions? I guess interesting to sort of see what we saw over the last couple of days. I mean, that's sort of forefront of our sort of future plans, which we'll talk about, a bit later as well. And Sam, are you going to talk to us about our cloud architecture? Thank you.

So what you see on the screen, we have spent a lot of time effectively looking at how we set up cloud infrastructure that help us in the bank. And we opted for a very, very complicated name. VPC one to VPC eight. Effectively, each one aligns to a certain business unit. So for example VPC one which is a fraud and smart disputes.

And we've got a couple of other processes. VPC two is on commercial, VPC five is on uh customer service for retail customers and so on and so forth. Um, when we build the cloud, we've started very, very small with one small VPC. Try to prove the concept. And then we expanded our cloud offering into other areas within the bank.

Also what we have is a lowest marketplace which effectively put all our APIs, all our integration in that marketplace area where any of these VPCs can benefit from somebody else's development. So if you build an integration to a certain area within the bank, you should be able to pick it up, even if you did not develop it and be very, very kind of it. Support you in terms of testing, development and improvements into your process. Also, what we did, we unified single sign on, um, within our uh, center of COE and the Guild. We managed to, um, agree a design principles into how would you get into Pega on the cloud? And everybody's using the same principles.

So what we intend to do or what we're doing now is build once and consume money. Um, just to give you some figures, 35, 35,000 users on these VPCs, we are doing approximately 600 changes a year. And also, um, the, uh, we have about 60 plus million cases a year. We consume we, we use in Pega. Over to you.

Um, well, actually, the second one, um, sorry. If you go to the second slide. Apologies. Um, this is a true kind of Center-out architecture that we've got. So what you've got, you've got Pega in the middle in fraud, for example.

And that applies to other areas within the bank. So you've got Pega in the middle. And we have integration in data sources for uh, within the bank like CBS integration which is your core banking systems, your, um, your, um, onboarding systems and other systems. On that side, you have the customers, uh, telephony and digital that also use Pega. We also have other areas, connection to schemes, and we have data sets, data mining and other stuff.

So we actually Pega sits in the middle and we orchestrate the work that we needed between APIs that we need to do the processing, data mining and data warehousing, and also connect to schemes within our infrastructure. Okay. Okay. So before we actually show you what we've what we've built, we'll just talk a little bit through our journey. So our journey actually started before um, the first date in this.

And really it started in about early 2020 with two kind of key events that really shaped our direction. So the first one was we had our old, uh, what was, I guess a, um, a flagship application, but it wasn't Pega six. So our old Pega six fraud application. So there was a conscious decision to rewrite that from scratch into Pega eight. And that was really a key enabler for us in terms of self- service, because we couldn't have done that by just migrating to Pega seven or lifting and shifting.

Um, the second thing was was really Covid and Covid proved that it was customer appetite for self-service. Um, and really what happened there was we went into lockdown in the UK. That was the 23rd of March 2020. And what we found very quickly was everything shut down, particularly travel and events. So there was a huge number of customers.

Um, we have a very, very good, uh, call center telephony, um, team. Um, but they were just swamped effectively. You know, all the customers were only being offered, um, you know, refunds. Sorry, not not refunds are being offered credit or rebooking. Um, and customers effectively wanted refunds.

So it was a huge amount of volume coming through. So we we put together very quickly a web form. And then as part of an MVP and again at that time we're on Pega six rather than Pega eight. Um, we put through a robotic solution which allowed us to, um, consume those cases in Pega, um, at least to allow our customers to be able to kind of get a recognition of the case, um, and then process it later. And what we found from that was that there was real demand.

You know, we'd done some research before and customers had have said that there was demand. Potentially for some of these types of self-service journeys. But we were a little bit worried that that that was the case. So as part of that, we decided to then invest in as part of our great journey once we deliver the the colleague view to invest in self-service view. So we did that.

And as you can see, that's sort of where our journey begins down there. In February 22nd we launched the debit dispute self-service. And what we looked at there was there were three categories that were really key in terms of volume. So it was goods and services not received. Um, it was travel and it was, um, continuous payment authorities. And that those three categories of I think there was about ten in total, um, were about 75% of the entire demand. Um, so that was what we focused on. And what we did then was we created an industry leading and award winning, um, self-service journey. And that basically gave straight through processing and included near instant refunds to the customer. We then launched the credit card disputes after we'd built the credit card journey up for the colleagues that went live in September 22nd.

And then by August 23rd, we'd reached quite a huge milestone. So we'd gone from zero self- service to 60% of our debit disputes that were being raised in the app. And of those, 65% of those were being straight through processed with no colleague intervention. And for those eligible cases, again, customers would get a refund in near real time. And one last point really was that also we were finding that the Representment rates, due to the quality of information and the the way we'd structured were actually lower than telephony rates as well.

So yeah. And I guess the sort of the journey from Feb 2022 supported the the mobile first strategy of the bank. So it's all done through three mobile customers could literally raise a claim within five minutes. Um, and as Elliott said, get a decision almost instantly. Um, it was interesting because at the time, we sort of we were still sort of falling out the back of Covid, and there were quite a few, um, sort of posts on Facebook where friends and family were saying, look, if you've got an account with Lloyds, try it if you want to make a claim, because I got my refund back within like 15 minutes.

So, um, we got some really good, good, good coverage. I guess to a certain extent, we became a victim of our own success because albeit I think we designed that journey mainly due to Covid, there was a real appetite in the bank to try and reduce cost to serve. So inevitably, there was a focus around how can we migrate more customers to the digital journey, create more opportunities to straight through process? Um, our customers clearly liked it. Um, so, um, there was an appetite to sort of effectively challenge us to do the same for, um, for fraud. So, um, in June 2024, and we used a lot of the learnings from the dispute journey, because you can imagine you do get people who probably come home from the pub on a on a Saturday night after having a few drinks and decide that they want to make a few spurious dispute claims.

So we saw a lot of that and we needed to design that out of the journey, certainly for disputes. Um, so a lot of the learnings from that went into fraud. Clearly fraud is more complicated. So we needed a lot of time to think about, um, improving triage processes and then really being thoughtful about what we wanted to straight through process, because clearly a very different decision for fraud than, um, from dispute. So we launched in June 2020 for for debit card only.

So this doesn't include sort of push payments and app. Um, we haven't delved into that. Um, that's certainly something that we could look at, um, at some point in the future. But we were really sort of surprised how quickly customers sort of got on board with it. I mean, when you launch a new journey, um, you know, we try and not go Big Bang because you never know whether the thing works and secondly, whether customers are like it.

So, um, we sort of launched it quietly on the app, but sort of within sort of 2 or 3 months, about half of all of our claims were coming through, coming through the app. Um, we, uh, effectively have reached a point at the end of August where, um, that 4000 claims is wrong. Actually, there's another zero that should be at the end of it. So we've taken 50,000 claims and our journey, which includes triage at the start. We're about 1.75 million customers have used that journey since we went live from June 2024. So, um, a lot of things to think about. Um, and clearly as well, there's some additional challenges coming through around, again, sort of reducing cost to serve. So um, that's I guess where we're at so far, we've got a short video that we thought would, uh, would show you, which was designed initially for colleagues. So again, I suppose we've got sort of 10,000 customers ring the bank every day, um, sort of speaking to us about fraud. So we've got to be really careful what we tell colleagues, because, number one, they might be worried about their jobs because customers are going to be using the digital channel.

But it's really important that we get their hearts and minds, because these are going to be the colleagues that will be speaking to the customers about the journey. Um, and again, we've had some great examples where customers have spoke to a colleague and the colleagues asked the customer if they've got the app and they've said yes, and they've actually taken the customer through the claim whilst they're on the phone, rather than transfer them through to a fraud agent. So I think where it works, it's great to give the colleague something visual rather than a piece of paper, which we know is a bit boring and nobody really wants to read train and documents on a daily basis. So let's have a quick look at the video. If you've noticed something that doesn't look right, it's now quick and easy to report fraud in the mobile banking app.

It only takes a few minutes and a few easy steps, and you could be reimbursed directly into your account. To start, simply tap the debit card transaction you want to query to view your options. What is it that has your alarm bells ringing? Is it the date or the name showing on your account? Some well-known companies look different on your account, so we've got a helpful list of company names that you can double check your transaction against. Next, we'll check whether you want to report a transaction as fraudulent or raise a dispute. Then you can continue to check any transactions you're worried about.

You can select up to ten transactions in total. We'll freeze your card so you know that no one can use it. Then we'll dig a little deeper on what's happened. So we'll ask you some questions, such as have you lost your card or Pin? Have you received any unusual text messages or shared your details with anyone? Simply answer yes or no. You'll then get a summary of the claim you're about to submit and we'll cancel your card.

This is to keep your account safe. Don't worry, you'll still be able to use your account and view your new card details in the app. Google Pay and Apple Pay will automatically update, so you don't have to wait for your card to arrive in the post to use it. And fraudsters can't access your money, but you still can. So what happens now? Well, you could be reimbursed in just a few hours, but if the claim needs further investigation, we'll pass you across to a member of our fraud team, either by phone or through an in-app message.

It's as easy as that. Okay, so that was the video. Um, hopefully you found that interesting. Thank you. So, um, how did we do it? I think the answer was obviously, um, the only way that you could reach, um, this results that we've just seen on the video is by implementing something that you could actually control on demand first time.

And as soon as you can. So you can serve as multi-channel. You could actually unify the business logic. So you build one central business logic. You can federate that business logic to the business to deal with it, but also you can provide the controls that you need at the front end.

So we don't really want to expose Pega to the front end at this point in time, because we were proving the concepts and we're working through an organizational and massive organization. So we have our own online digital journeys that is already available. So what we did is implemented something that would satisfy our kind of governance, satisfy our architecture and actually deliver transformation at the same time. So what we did, we utilized our um, which is our digital kind of platform, and we connected to them using the APIs. Um, and that allowed us to if we have any demand or any new scheme changes, we could just go and change that and it will auto reflects on the next customer click.

So that is the real benefit of delivering something of this nature using the EKS Okay. Thank you. So obviously now you've seen the video of what it looks like from the customer's perspective. Sam has explained roughly how it works from a DX perspective and how it hooks into Pega. Let's just step through the journey and just sort of show you how it works under the hood.

So we've kind of broken it down into eight steps just from a, you know, to explain it through to you. So the first step really is where the customer enters the journey. So the customer has got a couple of entry points so they can enter directly from the transaction on their digital statement, which is up to date in real time. Um, and effectively click that they don't recognize the transaction. It will start the journey from there.

Or they can go via the help function. Once they're in there, the next step is really around customer triage. There's two key points here or two key things we want to we want to kind of help the customer with. So the first thing is really to, um, surface other information about the transaction so often, you know, I think we've all had it before where you see something on your statement or you just don't recognize the, the The transaction. Um, so really here we're trying to surface more information.

So what other names does the merchant use? Also known as names. Um, the location of the merchant. Um, we map that and pin it on the map so that they can sort of look at that. Um, and Other the type of category that the merchant deals with. So that's one kind of key thing.

And the second key thing really is around whether it's actually a fraud or is it disputes. Now from a customer perspective, they often don't know the difference or don't care. Um, but effectively you will get people who will call in and say, you know, I've been defrauded by Amazon. You know, they've collected my Prime subscription. Well, it's actually no, they haven't defrauded you.

It's just that you didn't cancel it or you cancel. They still took it. So it's really around at that stage guiding the customer through. Is it really fraud or are we going to guide them through disputes journey, which obviously is also in self-service as well. And what's interesting there is at that point, um, around about 60% because of the amount of data and everything else that we provide the customer.

60% of customers that click, you know, help me with this transaction or I don't recognize it. Actually. Then exit the journey. So only 40% then continue on through to actually raise a fraud claim. So some of these, the rest of them go to disputes or they just drop out completely.

Um, the next step is really around. Um, then uh, allowing the customer to surface any other transactions. So if there's other transactions that they don't recognize, they can then add that to the claim. Um, and then the next thing really is around card free. So the customer research that showed that really in order for this journey to be successful, we had to give customers confidence as soon as possible in the journey that they, you know, their account was and the money were protected.

So what we do then is we, we, uh, call an API to visa. We freeze the card and we give the customer that reassurance right then that their money is protected before we do any further parts of the journey. Um, and that really has made a big difference in, And within the. It's all about retaining customer in the digital journey. I guess the real risk when we did the customer research was the customer is clearly anxious when they're raising a claim, and we need to make sure that they're given enough peace of mind to say, look, we've got your back.

You've told us that you might have a claim. We'll freeze your account. We tell them that they can unfreeze it themselves later on if they want to. Um, but it's definitely a differentiator to keep customers in that that digital channel. Yeah.

And at this stage, this is when we effectively start to create the case with, with Pega, we created a temporary case. Um, we go and retrieve the required questions from Pega. And then we start in the background, um, starting to enrich the case, enrich the details of that case. Um, and in terms of customer and everything else with data in parallel with the questions that we're going to look at on the next screen. Um, by starting to run some business rules.

Um, so what happens in is the front end then renders those, those questions. So in terms of the payload of the Rd And effectively does the lovely look and feel and everything else on top of it. We then. The customer that answers the questions and those questions, we've actually had to evolve those and craft those over time. And that's where DX API, as Emma said, as well, has come into its own in terms of being able to tailor the questions based on what we come, what we get through, to make sure that we're getting the best quality of data, to then map into the business rules as well.

Downstream. Um, the customer. Then in step six, um, gets the chance to review the claim. Um, and then once they submit that, that goes to back into Pega. Now, there are proprietary rules that we sort of run within our own there, but there's roughly around ten rules that we run at a high level.

And those are kind of looking at sort of four key areas really. So things like the likelihood of the claim being genuine. Um, back to your your point about people popping back from the pub and raising all their drinks. Um, claims back at the end of the night. Um, also things like, um, you know, the risk profile of the merchant, um, the transaction value and the count on that, um, and other other things like the maybe the, the type of transaction, is it customer not present, is it recurring, etc., etc..

Um, that's really where the magic starts to happen for the customer then. So effectively what then happens is um, one of two outcomes. So if we need to know more information, um, from the customer, maybe there's a couple of extra questions we need to ask. We then seamlessly hand that customer off, um, either to mobile messaging within the app, or we have a collect call within the app as well so they can go directly through. And obviously when they land with the colleague, they've got all the information they can continue to ask the questions.

They see where the where the business rules are, that the the customer fell out, um, and they can ask the relevant questions and then you could continue on with the case seamlessly from there. So it allows that, um, in the case of a 66. Successful STP, which is really what the customer really wants because they're all interested in getting their money back. What we do then is we cancel and reissue the card. So we've already frozen it previously, but we cancel it, reissue it, and it's reissued directly as well into the wallet so they're able to continue to use a digital card.

We submit the the to visa in this case because we're using visa for this journey rather than Visa or Mastercard. The fraud reporting chargeback via our Smart Disputes integrations. We refund the customer in near real time. So effectively the customer that can go and see that refund on their statement. Um, we send an SMS to the customer to confirm that, and we also then monitor any pending authorizations.

So if we've got any pending authorizations, we monitor those over the coming days. If they post, we then use some sort of business rules and logic on top of that to decide whether it's a genuine payment or transaction or not. And if it's not, then it gets attached to the case and the customer and colleague don't need to touch it so fully straight through processed. Thanks, Elliot. So, um, I suppose this slide is did our customers like the journey? Um, so a resounding yes.

I mean, the big number on the 18 out of 100, which I'm told is very good. Um, I guess that sort of stuff that sits underneath this is all of the customer research that we did. We were quite nervous because I think if you'd have gone back five years before Covid, we didn't really have the technology to do this. And I don't think you'd think that any customers would be comfortable raising a claim digitally. I think we would assume that they want to speak to a human being and be reassured that their account is frozen.

I think Covid taught us, actually, that customers have got used to raising claims digitally, um, and are more comfortable now than ever before. There's always the sort of far end of fraud, though, that absolutely customers do need to speak to somebody. So it's the customer's choice at the moment. We don't force the customer down this route. Um, and the odd examples that we've given, um, around the, you know, after hours sort of crazy claims after you've been to the pub, they would end up speaking to a colleague if they wanted to proceed with their claim.

Um, so that does happen. Um, but as I say, we weren't sure whether it would work or not. But I mean, the the experience benchmark at the bottom is, is is really high. Some of the comments that you'll see on the right hand side, and I suppose until you get direct feedback from customers, we assume this is what they might feel. Um, but you never know.

I mean, we have a process, I'm sure, similar to most UK banks where, um, if a customer rings the bank, they'll probably go through some sort of trial. So you'll speak to a colleague first. They'll try and see if they can help the customer. They decide whether it's genuine fraud or not. If it is, you transfer to another agent.

That agent then knows a little bit about the customer, but has to have a further conversation. You know, customers are regularly on the call for 40 minutes between sort of, sort of both, uh, both calls. Um, you can literally complete this claim within about five, uh, about five minutes if it's a straightforward claim. Um, so there's a huge difference between, um, customer experience on that and colleagues like it as well. I mean, as I say, the example that we gave earlier on around a colleague taking a customer through while we're on the phone, we could probably have never paid somebody to give us that feedback, but it was a it was a real opportunity for us to then share that with other colleagues as well.

In terms of the numbers on the next slide, Elliott, um, I mean, that that graph sort of says it all really. So we were really surprised that probably within 3 or 4 months, it did reach a level of maturity of around about 5,060% of customers using the digital journey. Um, STP rate is low at the moment. And that's by design, because clearly we need to make a strike a balance between allowing customers to get a claim and a refund where we feel appropriate. Um, but what we can't afford to do is sort of almost encourage sort of serial abuse of the journey because number one, it would cost the bank and secondly, would be sort of incorrectly tagging transactions as as fraud. Um, I guess, yeah. It would be remiss of me not to mention our award. So we've got a cabinet back at the ranch in Chester. Um, we've got three wards in there now, and these are the just two of them. So, um, great to be sort of nominated and winning awards for our user experience.

Um, but yeah, the the colleagues love a bit of kudos that, that that brings as well. Last couple of slides now. So I mean we thought we'd sort of share with you some of the lessons learned. I mean, I've already sort of touched on this, but don't assume you know what customers want and need. I mean, the design process was really important for us because we probably had about 15 or 16 customer listening sessions and we've got like a, a branch in Manchester with one of these two way mirrors.

So we get customers to look at different sort of, um, varieties of our journey and different designs that we've got. Customers do odd things that you just don't expect them to do. So You've got to make sure that that's designed in. Um, we eventually got to a great prototype, but I mean, you can imagine even now with live, we're changing it all the time because, you know, customers are hitting a dead end when we don't want them to. They're dropping out of the journey when we don't want them to do either.

So customer research is really key because, you know, customers need to want to use it and need to feel reassured on the journey. If you get that wrong, all you do is you just get customers confused and they just ring up the bank. So you probably find yourself answering more calls if you get that that one wrong. I think a key point as well was was really around in our landscape as well. Complexity was really large.

If you look at sort of card disputes or card fraud as a whole. So we've got credit and debit, we've got visa, Mastercard, we've got four different brands. Visa and Mastercard obviously work very differently as well. You've got fraud and disputes. You've got subtypes as well.

So there's a lot to go at. You can't eat the whole elephant in one go. So really we we wanted to start small. Um, as we said we started with the the Covid travel disputes. We then learned a lot of lessons through disputes, and we wanted to kind of learn from our previous journeys.

We did a lot of reuse in terms of integrations and rules, and try to kind of drive out that complexity in the process as well. Um, one of the kind of the key benefits I think we've mentioned before around DX API again was just the ability to tailor and the front end questions and really just tweak those to make sure that the customer is getting the exact data that we need, um, to map into the business rules and also critically, to map into smart disputes, so that we can then go out to visa and Mastercard and get a successful chargeback in one. That that whole process, by the way, is, is is straight through. So we go obviously from the form, through Pega, through the business rules, through through smart disputes, out to visa in real time and back through before we refund. So obviously that has to be right.

Um, otherwise you're going to have chargeback Sorry, representants or other things. And obviously the ability to tweak the tweak the front end questions then allows us to get better quality rules. And we can then also use the insights of PMI to then look at how those that's tracking over time and improving over time. So really key. And I think one of the other things that we've done from a technology point of view, we started with Pega Dqb1 as a POC and we learned a huge amount from it, especially around the integration with the digital front end channels.

What we did in the fraud journeys, we moved up one level where we actually saying, let's build flexible controls, where we can actually deliver a lot more closer, real on time services than what we had before, especially around the integration with the containerized environment that we've got with the front end digital self-service journeys and the digital channel that we've got. And the second side of things, we decided that we use the mobile devices only rather than a full digital offering. Yeah. And point number four on there is about building good friction into the journey. So this is I guess, the bit around you don't want too many customers just going to the end of the journey.

Clearly, a lot of customers use our journey because they don't recognize the transaction. That doesn't necessarily mean that they've got fraud. Um, so giving them rich information as you go through. So we consume lots of additional sort of statements and transaction data into Pega. So we're able to display that to the customer.

So we know that customers get confused between the data transaction and the data purchase. Um, there's regularly merchants who have got like a parent company that appear on statements, which confuses the life out of customers as well. So, um, you need to ensure that you've got enough friction in the journey so that customers don't find themselves hitting send at the end, when really we know that it's not a genuine claim. And I guess the last bit really is just collaboration across all functions in the bank. I mean, this is probably the most complicated journey that we've put live on the app since we launched the app.

Um, five different platforms all supported it. So you can imagine not just the technology that we've sort of talked to you about, but we probably had about 500 people across the bank sort of working on this at any one time. So being able to corral, um, the troops at the same time is a difficult job. But that's what made it successful as well. What's next? Oh, yeah.

What's next? Um, we've probably covered a few bits already. So continuous journey optimization. Um, just making sure that the, the user experience really sort of continues to be strong. We want it to be intuitive for customers because if it's not, they'll just ring the bank. As I've said, um, we've probably heard a lot about Process AI as well over the last couple of days, and we are sort of keen to just understand what opportunities we've got there.

Um, I guess some of you sort of spoke about some of the rules that we currently have, but we're keen to, um, take that one step further and sort of use some of the new technology out there that will almost predict whether a chargeback will or won't be successful. You can imagine the huge opportunities to improve STP rates. Um, with that, I guess the remainder really are just sort of our sort of, um, further evolvement of the journey. Um, one that we're really keen to build is almost to protect the bank against large fraud attacks. So, um, probably a lot of you in the room will be familiar with the autopay fraud attack that we had about 6 or 7 months ago, and the load of Tesla transactions.

It was a global thing. Um, I think we had 25,000 Tesla transactions in one day. So it drove loads of volumes, but it was metaphase. So it was a us, um, transaction on a UK account, which we knew would definitely be fraud. And they were like for one penny or two pennies.

So, um, if we'd have had this, uh, in place at the time and it would allow us to effectively white label all of the transactions and straight through process those in as fraud, It would save loads of challenges that we had around, um, sort of performance of, uh, of telephone answering and, um, sort of average handling time. Um, so not, not just gives a great experience but potentially protects you against bigger fraud attacks that might be, might be just around the corner. Um, and that's probably it from us all. So as a Bank of Scotland customer, I know it's a good app. I've used it not after going into the pub, but sure as well. It's really easy and at the end must you did touch upon it, but it must open other risks of fraud and then other self-service. People were worried about the jobs. Was there any big blockers or stakeholders that you had to overcome to get this through? Because I imagine people don't like change. Lloyds a big bank. Change sometimes can be hard.

Well, I mean, number one, there's always loads of governance, isn't there? So there's hoops to jump through. I think what helped us was the evolution of the disputes journey, because I think that gave us confidence that we can control the level of claims that we're happy to, to take through ESP. Um, as I said earlier, I think the STP rate is dialed down for all the right reasons in fraud, because if we think that there's potential abuse that's there, we just make sure that they speak to a messaging agent or speak to a voice colleague. We just don't take the claim. The real win, though, is that all of the data that the customers put in so far is saved because Pega has created an AI case.

So when the colleague goes in to talk to the customer, they'll bring up Pega. They'll see the data already, and instead of it taking them 15 minutes to gather data, it's already there in front of them. So it reduces the average handling time as well. Um, it's fair to say that I think there's still a bit of nervousness out there because it's new technology. Everyone's a bit scared that 66% of customers are using it already because, you know, that's a big shift to digital.

But I think we've seen with the feedback and we've seen the controls that are in place, that it's given the bank confidence to sort of continue to evolve the journey as well at the same time. Thank you. Anyone else for questions? If you go down to the microphone in the front. Thank you. All right.

Great presentation guys. Really enjoyed that. Um how did you see any impacts on complaints raised by customers post raising the dispute. The the volumes come down or the nature complaints changed. It's really interesting.

So, um, yeah, we had to present to the board about sort of a month or so ago on this because, I mean, historically, customers were complaining that they didn't feel like they could raise a fraud claim in their channel of choice. Um, and this is only for debit cards only at the moment. So it's not been expanded to any others. Um, but the complaint profile has definitely reduced, obviously, as well, because more customers are using digital than fewer customers are waiting because it takes a bit of time for us to, I guess, sort of reduce the number of colleagues that answer the call. So definitely it's upside on complaints.

Oddly enough, though, you've now got some customers that are complaining that they can do one thing on the app, but they can't do another. So Apps and Scams is a is a big challenge for fraud. That probably is our next opportunity. But nevertheless it's yeah a difficult one to to to to work through. Follow up question as well.

Yeah. Any plans to do anything with voice. That's very app based what you've presented today. But any voice conversational type interfaces on the horizon. Absolutely.

Yeah. And I think again it's just making sure that we don't confuse customers too much. I mean, without going into detail, there's some new technology which will allow us to link into some of the, um, sort of, I guess, uh, sort of guardrails that we've set for the, for the digital journey and take advantage of that through, through voice. Um, and with all the exciting stuff that we saw around Agentic AI, I'm sure you guys at Nationwide are keen to take advantage of that as well. Thank you. Yeah. All right. Good afternoon. I have a few questions. Actually, let me start with reclassification.

So was there you only do debit cards? I just heard, right? Not credit cards. Correct. At the moment. For some folks that came in the dispute side and actually it was fraud or vice versa. And you had to reclassify.

Yeah, absolutely. Um, what percentage was that? Um, we probably had about 30% of customers that started in the fraud journey, but because of, let's say, clever triage, we then moved them into the other journey. And that is really key because number one, customers don't have chargeback rights if they go down the wrong journey. And potentially the bank is obviously, um, liable for any, any write off costs. Okay.

That was my second question on chargeback rights. So do you automatically decline or do you accept the claim and then decline it later? With fraud we always allow the customer. Yeah. If we don't accept it in the digital journey, then the colleague will need to either accept it or decline. So we don't have an STP rule.

Let's say that declines the customer now. Did you see? We have not gone online yet because we're nervous that it's going to increase volume in general. So as my EVP. Right. So did you see an increase in claims either on the dispute side or the fraud side? Yes.

So we saw some increase. Um, I mean, it's interesting because I think Elliott referred to it before around the repossessed rate that represent rate actually went down. So probably what we captured was more genuine disputes or genuine fraud, because it was a bit easier for the customer to raise a claim. Dare I say it, a customer who maybe has got a 4.99 pound transaction probably wouldn't have rang the bank. Um, because it's 4.99 pounds, because it's digital.

They can. I think what we're cognizant of, though, is that if we can straight through process that journey without human intervention, and it's not impacting our present numbers. Then it's a bit of a win win for the customer and the the bank. But if you don't have chargeback rights, it's going to increase your fraud losses, right? So did you see an increase in fraud losses? No, we haven't seen an increase in fraud. Losses.

Because that looks like a 50% or two X increase in your claims, right? I think that's volume as well. So to to to your point, a lot of volume is very low value. So you know you know there's a cost to serve isn't there. If with a telephone call. Um, and also you know, the cost of raising that.

So I think, you know, that's one. OK I have several others, but thank you. Okay. Thank you. Delighted.

Yeah. Hi. Hello. First of all, I'm glad we flip back to this slide because I was going to ask about this. I am amazed by mid-July to less than six months later, you saw over 60% of the self-service case volumes just magically appear.

How did that client, Education or customer awareness process go. Whether it was marketing, in- app notifications or other. Yeah. I mean, you almost don't want to market this journey for all the right reasons here. It probably happened because, I mean, typically our experience on the app anyway, customers, if they're unsure about a transaction or they need help, there's one button that effectively does it all.

So first of all, it'll maybe show you a map. Then it will give you more, richer information about the transaction before we automated the fraud journey in the app. Customers effectively pressed a button to say they thought it was fraud, and they were encouraged to click to call or speak to a messaging agent. So effectively, what we've done is we knew that there were thousands of clicks. Let's say every day to those.

So customers are still clicking, but rather than clicking click to call or messaging the starting our journey. So I think that's why we saw quite a big uptake. This this wasn't a brand new journey where customers never had the option to raise a query that was already there. So we probably took the benefits from that. Not great.

Thanks. Okay. Thanks, everyone. One more, one more. First of all, congratulations.

I think it was a great example of how to leverage the API. Uh, for that, it's. Congratulations. But I just have a question. Whether you use the constellation SDK or it is just API version two, uh, rendering the UI, and then you're creating a fraud case or it is a fraud case.

Good question. I kept it as vague as possible for a very good reason, because if you are kind of working in the PegaWorld, you would almost certainly went through this kind of due diligence trying to understand what is the best solution for us. Um, obviously the SDK is one of the best things that you could ever do. However, with an organization of our size, we have four major brands. We do 27.

Is it 27 million customers? 28. Um, it's a bit more. It's a it's a constant kind of discussion between us and Pega that sets and orchestrates and the front end technology that we have where they use the they have their own kind of, um, digital platform. So what we wanted to do is to deliver uninterrupted, um, kind of, um, technology. So what that means is effectively we want to deliver, but we want to use what we've got as well.

So what we what we did, we have built we have used what is available at the front end, but we have built our Pega library within the um, within the, the, the digital journeys as well. So we opted for an optimized DX lbg design that we are satisfied to accept and support. Okay, and that's a wrap..