PegaWorld iNspire 2024: Building Customer Trust: Turbocharging Early Success with Pega Customer Decision Hub

So it's been about 30 minutes going through a couple of slides just to give you a sense of the story. Try and hit some of the highlights of course with with with presentations like these, you're never going to cover everything in 30 minutes as we've been on this journey for a couple of years. So your your Q&A at the end, your questions at the end would be really important just to kind of connect the two stories together. Um, I am going to take any of your questions. I'm going to answer the easy ones myself. And then I'm going to pass the more difficult ones to two of my team that are in the room with Alexandre, who's our head of decisioning, and Paul Kelly. I see him slinking down the back there as our product owner for our customer engagement engine. So let's get going. Just a little bit of scene setting and introductions at the start.

So this is a one page overview of me and my career. About 30 years industry experience across all of the different companies you see along the top of that page there. Um, in the industries you see next down in the countries just below that, and then in the different roles, uh, as I've gone that journey over the last, um, almost 30 years as it is now, ten years of Bank of Ireland, and you could chunk down the Bank of Ireland's story or my, my time in the bank into the three pieces at the bottom of that, that slide. It's been the first three and a half years managing one of Bank of Ireland's regions, one of its four regions in Ireland, retail and business banking. Uh, very good grounding for for um some of the further roles I've had come out of that to um, lead two projects that were going astray, one of which was GDPR in 2017, 2018 and time for Go Live. And then I've been part of the newly formed group marketing function since 2018, um, initially marketing transformation. And that's where we connected with with Pega and the CDH deployment, but over the last three years, I've taken responsibility for group customer analytics, and you'll see a lot of the references to customer analytics throughout the course of the next 30 minutes in the presentation. The one point I would make, um, for a qualification in engineering and for most of my career, I had left that qualification behind. I am finding more and more over the last two years, uh, unsurprisingly, with the explosion of GenAI, um, that it's not just AI and ML and all that, but now with Gen AI, that qualification is really coming to the fore in terms of being able to understand the pace of change, what's going on, how we can factor into what we do.

So that's been been a really interesting I had thought I'd left that engineering degree behind, but I have not, um, Bank of Ireland, 241 years old this year. The building with the pillars on the right hand side there is our college green branch. I put up this picture just to remind myself, as I as I go through some of these talks about the legacy, the history of Bank of Ireland, um, our position in Ireland, our responsibility to not only our customers but to the country in terms of, uh, the financial well-being, uh, position that we lead. Um, and also to remind ourselves of the amount of data that we have, um, and that we're custodians of for our customers. It's a it's a really good reminder. Two quick, fun facts on this picture. The Bank of Ireland branch there in College Green is just across. For any of you who've been in Dublin from the entrance to, uh, the Trinity College entrance. Um, it's the oldest, I'm told, still in operation.

Banking hall in the world. Fun fact one. Fun fact two is the building to the left of it there with the on the corner is now a Starbucks. So if you find yourself in Dublin, visit Trinity, come across and meet us in the oldest in Operation Banking Hall and then go for a beverage. There are a couple of pubs down that street. If you find yourself there a little bit longer For a more modern take, let's say on the Bank of Ireland group, there are a number of different subsidiaries, and the breakdown of these subsidiaries and the market dynamics is really important. We have over 4 million customers in Ireland and the UK. There's about 3 million in Ireland, Republic of Ireland and about a million between Northern Ireland and the rest of the UK. Business retail Ireland, typical retail bank, all of the normal products you'd expect, all of the normal products you'd expect and with white label some others.

Retail UK is largely a partnership bank to the Post Office Network or New Ireland wealth and life insurance business. Recently we've bought um, Davy Stockbrokers and that picture belies if any of you know Ireland, there's been a lot of of change and flux in the financial services landscape in Ireland. Over the last couple of years, we've had two traditional banks have left the market, Ulster Bank and KBC, which has been a big opportunity for us that we've been able to to capitalize on with some of what we're going to talk about today. But we do have a very vibrant fintech landscape as well. Um, and a lot of fintech competitors, um, at our heels, uh, for, for our customers, that's Bank of Ireland. Uh, yeah. Look, we've talked about the very different market dynamics and how we and the customer, um, analytics community and group marketing overall, uh, take all of those different market dynamics into account when we're creating campaigns and communicating with our customers. Let me just talk a little bit about the customer analytics team. We'll leave aside the marketing transformation piece.

Uh, and I'm going to talk about this as it has been really important in terms of of how we've gotten some success when we, um, took on the customer analytics function three years ago, we had a couple of different challenges. We had, um, a very siloed team across those four different Subteams commercial insights, data science. Our decisioning factors, of course, was brand new at the time a marketing effectiveness so siloed approach, uh, no real proper understanding amongst our our business stakeholders of what the difference between those teams were. And that was driving more of that siloed approach because our business stakeholders were coming into any and all parts, uh, asking for things to be done. We also had high churn, which is a function of what was going on in the marketplace at the time and, and low, um, engagement, all of which, um, together can cause your problem. So we've set about myself and the management team over the last couple of years of breaking down the silos between those subteams. First of all, we have, um, increased engagement from 50% to 80 odd percent. Let's be quite successful. We've eliminated, um, almost churn 3% last year.

Um, and we've done a lot of work on virtualizing these teams such that the Commercial Insights team, which is which is what we would call front of house or face offs to the business, uh, has a virtual team behind it. A member of data science decisioning, marketing effectiveness and so on. And we found that to be really successful. But the last piece of the puzzle, um, to, to, um, to get the business stakeholders on board, it's very hard for a financial, some financial services background, a banker to understand what you're talking about when you talk about the difference between commercial insights and data science. So we come up with a restaurant as an analogy to tell that story. And it's been it's been hugely successful. Commercial insights are, as I said, our front of house. So we we align them to, uh, to the maitre d in the restaurant analogy, our data science team, traditional propensity model builders, we align them to the to the chefs in the kitchen, in the restaurant, um, our decisioning team, you've got a sushi bar in the middle of the restaurant. You think about all of those impressions to customers and customers choosing whether to engage or not.

We make that similar to a sushi bar and a customer deciding whether it's going to take a tray off or not. And then last but not least in the story is our marketing effectiveness. Um, our food critic in the restaurant. And that story has really gone down well, and it has stopped the business stakeholders from going into all different parts, knowing exactly using that restaurant analogy where they might come in. One of the follow on questions we get then is, but how do you relate to the data office? And we say the truck at the back door providing the fresh raw ingredients to the restaurant. That's the same as as a relationship that we have with the data office. We want the data when we need it. We want to collect it.

We want it fresh. And that story has gone out really well. Just the last piece around the team. Um, again, part of of how we transform the team for success with, with the Pega deployment over the last couple of years is we in the past, we were just operations. The team were just doing operations. I think part of the reason we had such high churn was we were burning people out, and we weren't allowing them to think about their futures or leverage some of the exciting things they had done. A lot of people coming out of college and coming into this sort of environment and then being thrust into operations only. We've moved to you have three different things that you do all the time operations, transformation and experimentation operations. About 85% of what you do is the bread and butter.

It's what you do this year. For this year, transformation is about 10% of what they do. It's this year, but it's not for this year. It's for next year and the year after and then experimentation. It's largely proof of concepts or areas of interest. There's about 5% of what the team does and the things they're doing in experimentation this year. We don't expect to come to fruition in operations for 3 to 5 years now. If it's sooner, that's great. But that's the level of expectation we set.

And you can see from the right hand side of that slide if we get the team working, first of all, when people come through with experimentation ideas, you start to see your leaders of the future. Already you're starting to think about what they're thinking. Like not everyone comes comes forward with projects, but those that do. You start to see a different side of them than you would do in their in their normal operational lives. But if they come through with ideas to solve some of the operational problems, those become your transformation items for the future and then become operational. So you've got this, this nice positive cycle at the right hand side of that of that screen there on the slide here we've got introduced myself, induced the bank and introduced a team. So let's get in a little bit to the why, the how and the what of of CDH or our customer engagement engine. So we um, if you go back to what I said about we've got over 4 million customers, we've got a very diverse product set. And prior to Pega, um, we had no ability to orchestrate messages across the different channels.

We had no consistency of message. If we wanted to send a message via the app, it went to all customers. Over 3 million customers. Got it. Um, we also have a very digitally active customer base. Uh, a huge percentage of our transactions are via the app or on digital channels. And we had no way to to pull all of that together. And so it was clear to us, uh, from early that we needed to get on the path of, of a decisioning engine. And then that also helped us with the the challenge you're seeing on the screen here, which is we we knew we needed to move from reacting only to our customers to being more proactive and ultimately are ideally preemptive.

And what we mean by that is if we take an example, customers have moments of need. You'll see on the left hand side center of that slide there they have a moment of need. And if you draw a timeline horizontally to the right of that moment of need, let's call it a change of address. Customer has a change of address. Have they changed address or are they planning to change even those two different states? As you think about it, have different levels of urgency as to how we should respond to the customer. But the customer decides at what point along that timeline they're going to come, and at what channel they're going to come into us in. We then have to have the right service levels, as you'll all be aware, to maintain the loyalty that we want so that we don't start to hit churn issues. But what if we could move from being reactive to being proactive and sticking with the change of address?

One of the very first MBAs that we brought live was a change of address NBA reasonably simple. We looked at all of our address data and if certain criteria existed, we thought that was potentially a data issue in the address field. And so we presented an NBA to the customer proactively. Um, we went from average 500 change of address requests a day to 11,000 on the first day. So hugely positive in terms of of customer present the address. Is that the right one? If not, change it. Now we have to turn that one off and throttle it before we brought it back to life. But with huge learnings from that, not only in the NBA itself, but also because we had such an explosion from that one NBA we got huge stakeholder buy in around understanding what this decisioning engine was and how it worked.

Just just one NBA on the very first day we got, we got massive understanding and buy in because of that, that we wouldn't have gotten, um, or would have taken us longer to get. Ultimately, we want to become more preemptive against sticking with change of address and thinking about us. Mortgage customer draws down a mortgage that's not an investment mortgage. Guess what they're going to be changing address. So how do we bring that to life for them in a way that's more service oriented in terms of of um, that was the why the the how um, we broke our project down. Uh, the initial MLP one minimum lovable product, one into three areas. We had the technical go live that we did with our IT colleagues and with our partners at RBC. We had all of the content and the NBA. And then last but not least, the third area was around Optmodel.

And we had we had a decent optmodel prior to this, but we did work with our partners in merkel to adjust the optmodel as it was because the bank prior to this was very siloed. The campaigns were very, very product driven, very siloed in terms of of being product driven. And typically in a product driven environment like that, it's the it's the customer. It's the it's the products with the highest commercial value or the loudest voices that will get the, the, the, the most traction. We wanted to move away from that to be more needs based, more customer focused. Uh, and so we have been moving. I won't say we're there yet, but we have been moving towards what you'll see here in the screen for the three customer states of find, win, grow, serve and protect. Uh, and there's a very different tool set and mindset. Go with the squads as they operate across those three different customer states.

Find when being brand new to Bank. Uh, not not another product to an existing customer that's in grow serve. Uh, and if if um, we had back at the start actually we have referenced down in the bottom right hand corner of that screen our strategic pillars. And we do leverage the group strategic pillars to tell this story and continue. One of those is about stronger relationships with our customers or deepening customer relationships. And we're able to to bring that strategic pillar to life. Uh, in terms of, of of how that model, um, has worked. When we look at, we talked about the why, the how and the what, and this is our sort of simple overview. I'm not going to put an architecture slide up on the screen, but our simple overview of what we have, uh, very simply on the left hand side, uh, is, is all of the data that we have fed into, um, to Pega and we continue to ingest that, in fact, uh, in our most recent upgrade, we know we've got data that goes goes into Pega.

But then there's always another campaign that will come along that we don't have the data for. So even enhancements like having an ad hoc campaign table that we can add more data at very short notice on the left hand side as being as being critical. We're working with our our data colleagues in the data office. Um, they're deploying an operational data store at the moment, and we will get live data feeds and event triggers from that, which will be hugely helpful to us. We talk about the role of analytics team next across on that page and the role they play in creating the propensity models, the predictors of which become the primers for the adaptive models in CDH. And so data science team play an important role in this overall picture. Before we feed into the brain and then out into the to the channels on the right hand side, we have um, our initial ambition was ten channels were five in, uh we're on we're just moving on to our sixth. But there are more than ten now that have materialized since then. And you can see roughly, we spent the first two years with app and desktop only, and most recently we've added um, email, SMS and push.

So we're looking forward now to getting those live. Getting more traction with them, getting more campaigns through those new channels, and then seeing what that's going to do to our CTR as we progress through that. We are excited about the upgrade to, um, to version 24, and we are looking now at what I can do for us both in terms of the on demand recommendations and the and the artifact generation as well, uh, as well as, uh, some of the new functionality in the Strategy Optimizer bundle, 1:1 Ops manager and so forth as well. So there's a lot going on in that picture. Um, and the team have done a super job in bringing that to life for the marketing community and for our customers. We talk about we work through all of that, the, the, the why, the how and the what, of course. measurement and the results are critical as well. Um, so in our in our measurement framework, we've tried to bring together all of the different KPIs that are floating around at the bank. And this is the sort of first reference to that.

Again, maintaining that view to the customer. States find win, go serve and protect and then picking a selection of KPIs. But what we found in the bank is that there are particular people that have pet, pet, um, KPIs, things that that they believe are the most important KPI. And for us, um, putting those KPIs on the marketing funnel, showing where those KPIs operate in terms of the different channels, has been really important to get everyone to understand that it's a suite of KPIs, not only 1 or 2, not just there ones that are important, but this full suite of, uh, of KPIs is important. And where it sits on the funnel will determine different things. Uh, so that's been, um, been really positive for us and that's then laddered up into, um, this view. Uh, and again, going back to our strategic pillars and particularly deepening relationships. What is the, uh, the objective we're trying to achieve? What are the dimensions to that objective?

What are the KPIs and what are the levers such that if something on that page goes from green to amber or red, we start to understand what the what the underlying, uh, levers are, rather than just reacting to the thing that's gone amber or red. Um, and this is a huge piece of work that's going on, um, since the middle of last year to try and corral all of those KPIs into that, uh, measurement framework and bring it to life for all of the different stakeholders in the bank. Some of the other results you'll see top left, the growth in the library of MBAs over the last couple of years, since we started, uh, 23 versus 22, we doubled the number of MBAs we have, and we continue to have good demand for those. And again, um, allowing those to come to life in the new channels will be really important. We did start off with a strong focus on service, uh, MBAs rather than sales or growth MBAs. Um, while there were some in there, it was largely service dominated for for the first, uh, 6 to 8 months until we got established and understood what we were working with. We have a good spread of MBAs across the different business areas. You see, uh, bottom left there as well, and particularly categories on the right hand side. Um, fraud has been a big area for us over the last couple of years.

Um, and we've had, uh, last year with a million engagements with fraud MBAs, as we've put out a lot of educational, uh, messages to our customers to try and help them understand what they should or shouldn't do. We're seeing that filtered through as well in terms of, um, success in other parts of the business, like our suspicious transactions line has got quite a sizable reduction in call volume coming through it, and that reduction has maintained, because we see a corresponding increase in self service for those customers that have stopped calling and are now self-serving anytime they have a fraud related transaction. So that's been really important. Similar stories across many of the rest of the service, MBAs and the sales NBA have performed quite well as well. In fact, this picture will show you the sort of 5050 split we have between sales and service. Um, and we do retire MBAs. Uh, we do we do look at the MBAs critically, the performance of them, retirement return if they're if they're no longer in use. Um, and the focus we've done really well with the with the service MBAs, the focus has really been to try and promote service and nurture messages and give ourselves license to have that, that conversation with customers when the sales opportunity does come. And we have some good success on lines like personal lending.

In fact, when we look at the response or the or the click through rate for customers and the subsequent product purchase or draw down at 30, 60, 90 and 120 days, we have very good attribution throughout that that timeline as we continue to track the customer's behavior. And so that extended timeline has been important as well. In terms of learnings. Um, on the deployment. So going Pega Cloud, I think was was really important for us. We launched with, uh, on two channels with version eight, five upgraded eight, eight, uh, pretty seamlessly as well. Um, important that we went on the inbound authenticated channel. It left us less challenged. Managing things like GDPR compliant consents.

So Pega Cloud inbound authenticated channels is important. And then getting the structure of the the set up right for those three pieces, the technical deployment, the op model and the content and MBAs. Um, that that we had set up in terms of what's next, um, for us, more data, more MBAs, um, and more channels would be the first piece we will upgrade to version 24 in the next couple of months, and really looking forward to what's going to come in terms of the GenAI components. And we're starting to investigate that in the in our sampler area as well as we think about how we bring some of the artifact generation on on demand recommendations to life. That's it. Any questions? Thank you. Thank you Joe. That was a really insightful session.

If I can just ask you to if you have a question, just step up to one of the Q&A mics and ask your question. Thank you. Thank you so much. This is the offer from citizens Bank. I have a few questions around the propensity score. So you said you were using NBA runs for growth and service, correct? Right. So we also do a service as a communication solutions as well. And I'm curious that do you use propensity scores for both of them and if so, do they actually share the omnichannel approach or the NBA layers are distinct between service and growth solutions.

As propensity scores for for growth or service propensity scores are, propensity models are largely growth related. So um, acquisition or retention rather than service? Service are more, as I said, around things like fraud or educational. It could be financial well-being. Uh, we've had a lot of success with on the service side, with financial well-being, health checks and then and then more assisting the customer. So we have a decent propensity scores on the service side. Thank you. And also for just to stay on the growth side, um, do you use like adaptive model or you leave everything to the data? Scientists come up with the propensity score on their machine learning models and use a rule based approach.

We use the, um, we use the adaptive models in Pega. So our the predictors or the or the scores of propensity models, we use them to feed the adaptive models to prime them and let the adaptive model take over. Then we do very little else with the adaptive models uh, than prime them with propensity. So like, you know, the data scientists team, like, are they providing any in-house solutions to this or it's more like, you know, they're designing on which, you know, AI model that they should be using purely on prediction studio. They're they're designing all of the all the models. Yeah. Data scientists are. Yeah, yeah. We're doing a little bit of work with a company called H2O at the moment to try and accelerate the pace of development of a propensity models and try and, um, you know, we have we have about 40 propensity models, um, at the moment.

But we believe we need to get into that, you know, well over 100 models. And so we're looking for ways to accelerate that with, with a platform called HCL. Thank you. And I don't want to take anybody else's time. And do you use the ops manager as well like with the NBA designer. Because I heard, you know, the ops manager is more like revision management, right. And how effectively actually using the NBA designer, because usually a lot of time we use the Customer Decision Hub strategies. Yeah. But you know, maybe like a good idea would be how many actions are you able to do developers are able to create.

So we've just we've just gone live with one ops manager. And we're we're really excited about that because it starts to take the CDH out of our hands in the analytics team and give it to the rest of the marketeers. And so we've just gone live in the last little while, so we're rolling that out. But it's really I think it's really important to get as many of the different parts of the marketing community involved in using the platform as possible. Um, so that's that's a recent, but we're expecting we're seeing some good traction on it and expecting some good success. Thanks so much. You're welcome. Hi. Thank you so much.

I have a question. You had one slide with the kind of segments on the side when you send these out. And over time did you, um, collect how call it customer segmentations. You know, older people, younger people, people with more money, less money, more accounts, accounts. How they react, how they perceive, how they're, you know, you're sending it out, but you also want to get there. On the action taken on their end. Yeah, we do. Um, so we we do a lot of work across the different segments for the different MBAs to To understand exactly what the attribution is for the different actions. There's a lot of work in our marketing effectiveness team, and that's what we're trying to bring to life in the measurement framework, so that we can see across the customer states of find, win, grow, serve and protect.

But then for the segments as well, because, you know, we have a natural inclination in our find win towards the youth accounts, because we know exactly what the bell curve of customers, who we get in at that, that youth stage all the way through their financial lives. But we do track the rest of the, um, the segments as well, and we evolve some of those segments. So for example, Premier is a segment that's not age driven, it's more wealth driven. So we have different, um, you know, you've got vulnerable customers that are in and out of vulnerability. So we do have ways to track the response of different segments to different NBA. Thank you. Hi. Thanks. I had a question more related in how is your relationship with the business objectives?

How do you translate the business objectives and transform it into an NBA. What's the process? You you have to include that and also continue it towards the end and the measure of it. Yeah. Good question. So we I talked about the model that we put in place, um, as part of the original group marketing. And then we evolved that model as part of the deployment of, of CDH. So we set up these cross-functional marketing squads that were underpinned with the different CEOs in marketing. And so we have our commercial insights team for my area is a member of those squads that cross-functional squads.

So they play a role in translating the marketing objectives into analytics. But then we also have that team that interfaces back to the business. And so we get this triangle effect that happens between the business marketing squads and our analytics. So that we think this link back to the business from analytics is important because we are educating and providing insight. Data dashboards all the time, and it's often those insights that they use to drive their marketing demand. So it's true that um, that what I would say triangle. And which of those three parts of the triangle is the one that executes the NBA. Uh, the next part in, in the, in the analytics community. So the on the analytics side, it comes in then to our data science community, if there's a propensity model needed, and then into the decisioning practice for the, um, NBA to be programed and executed via the brain.

Thank you. No problem. Hey, Joe. Uh, Rusty Warner with Forrester. I've actually been in the bank in Ireland. I still use a photo of your customer moments in some of my presentations. So thank you for that. Uh, but my question is also on measurement, but more about how are you looking at customer outcomes and financial well-being and trust and those kinds of customer oriented metrics, and translating them into the kind of business value that you can then build into your models? Yeah, it it depends on, um, on the different areas of financial well-being.

We, we we would say we're number one for financial well-being in Ireland. If you just stick with that one. And we have a series of KPIs to support, uh, financial well-being where and it's wider than than what we do. Where we look at it is some of the MBAs that we have in the service side are are educational for customers and prompt customers to go through a financial wellbeing health check and the the follow on for that then is if if, if those customers do have financial wellbeing financial needs review. So we can track the presentation of of the um, the message, the engagement with it and then all the way through that journey. And we would do similar for, for, for different areas of the business, but it's trying to understand that journey and how you relate the different metrics to it. Again, I keep coming back to the measurement framework because it's within that that all metrics sit and that one starts to extend itself out into things like NPS, SES. We use our NPS and NPS, um, and then brand awareness as well. And so we have a whole suite of metrics to, to align to those.

But it is very much business, business driven. Um, financial wellbeing is one that that goes across a lot of the different business areas. Does that answer the question? Thank you. Any further questions for Joe? No. Well thank you Joe. That was a fascinating presentation. Um, really good.

And thank you for traveling all this way to be with us. Thank you. Thank you, everyone. Thanks for your time.