PegaWorld 2025: Using Pega's AI-driven Platform to Innovate at Bradesco

PegaWorld 2025 - Using Pega's AI-Driven Platform to Innovate at Bradesco

Welcome and hello. My name is Michael Baldo from Pegasystems Industry Principal for lending and all that parts. I'm very happy to introduce two very, very intelligent and powerful ladies, Rubia and Mariana from Bradesco in Brazil. They will tell us about their stories and their journeys, how to improve wholesale banking, and how to actually get a lot of value out of these discussions. Please. Rubia Mariana.

Thank you. Michael. Good afternoon, everybody. I'm Rubia and she is Mariana. Okay. So thank you very much for coming. Let me before we start, just ask who here is from Brazil or has been to Brazil. Can you please raise your hand? Oh, great. The the majority, I would say.

Well, even for those who haven't been to Brazil, I'm sure I don't need to say that we are a massive country. And Brazil itself, it's bigger than the entire European Union together. Okay. And it's not just that we are big in terms of territory. We are complex, we are diverse, but we also have a lot of opportunities, thank God. Right. And Bradesco cone sits at the heart of this landscape.

Okay. We let me put here. We are one of the three largest banks in Brazil. We operate across all segments. We are a full service bank. We service individuals, small businesses, up to large corporations and institutional clients. And we also serve them abroad. Okay. To ensure the connectivity around the globe, we have over 73 million clients. We are we have a very strong brand. Our brand is top of mind for 18% of our population. Okay. We have 215 million US citizens in Brazil. So this means like over being top of mind for over or for circa 40 million people, which is a lot. Okay. And we are top two private banking, top two in affluent market. So we really have a very strong operation.

Okay. Nonetheless, as many other companies in the from different industries we suffered with the Covid pandemic. Okay. Our results suffered. So early last year, we launched a transformation program aiming at achieving sustainable recovery of profitability. The transformation program was based or was structured around ten pillars, with credit being one of the main, not only because of the financial potential involved in the credit cycle, but also because of its complexity and the opportunity to transform, to improve OK.

And here is where Mariana and I. Joined the program in the Wholesale Credit Analysis Transformation OK. Our main goal was was to reduce the operational risk, to improve the decision making, and to decrease the time to yes, time to yes in the wholesale business. It's crucial it's any segment for the banking industry, but wholesale, we have other customers that have high expectations okay. And the window for the deals it's very short. So whoever gets there first increases the chances of closing the deal. So we know that we've measured that. Okay. And you even are able to price the deal better if you respond fast. So time is of essence for us.

And it's a very complex business. Right. Because when you look into a company, it can spread into or to a group. It can spread into many companies, many products like we have over 120 products and we have a portfolio over $60 billion. We have over 25,000 companies operating with us. We have different guarantees, we have different documentations, we have different, uh, or many layers of approval. So it's a complex business and we handle over 3.5 Point 5000 credit analysis per month, with a circa 200 people involved in the process in the credit department.

Okay. So, uh, what were the main challenges that we had to deal with in this process? Okay. We engage with people from the entire bank like representatives of different areas and, uh, segment operations, legal, uh, credit itself. We gather everybody around and discussed what were the challenges, what we wanted to achieve. Okay.

So we have, uh, we had now we have Pega we had an not user friendly, uh, front end. We're going to show you here. Our front end. Okay. And we had 55 linear and manual steps in the process. So very fragmented. And this came because of the many acquisitions that the bank has done, uh, along his history. We have over 80 years of existence. We bought many banks. Okay. So we have like 100 legacy systems. So it's it's a complex infrastructure.

40% of the time of the analysts was spent gathering data. They were not using, like, the intelligence to analyze and and to contribute to the process. It was just gathering the information. We lacked sufficient structured data even for the next steps that we we want to use. GenAI. You need to have structured data and very complex Architecture. Don't comment on that. So the complex architecture meant that we spent more to maintain and also to evolve. So it was harder to implement new journeys or new platforms that were more user friendly and also more efficient.

Thank you Mariana. She's the intelligent one. I'm just the the speaker and and as she said, high maintenance and high development costs. So those were the challenges we're facing. We were facing in our credit approval process to be what we wanted. One stop shop with with a with a friendly front end, automated steps and decisions, AI insights to guide the credit analysts with suggested decisions and even to organize our our. Um, uh, our team OK flexible architecture. You need to be able to to to change fast, right? Our market is evolving very it's constantly evolving. We need to be able to be to adjust faster and 100%. Uh, cloud hosted. So that's what we were looking for.

But obviously this transformation comes in phases. You can't do it in waves. You can't do all this at once. Okay. And then I'll let Mariana again explain the difficult part, how we did the the waves.

Okay, guys. So as mentioned, we constructed a transformation in waves in three waves in the first wave, which we just recently concluded, we rebuilt the workflow from the ground up. So we constructed an intuitive interface. We also simplified steps and allowed our analysts to work in parallel. So before analysts had to wait in order for a technical Allianz to do their jobs. However, now they can work simultaneously and input the information in a new workflow.

We also allowed for attachments. Yes, the last system did not allow for that, and now they are able to attach information regarding the companies or the transactions that are analyzing directly into the proposal. We also structured data, as Rubio mentioned, that was essential for our next waves. So we built a strong foundation in order to be able to move on to GenAI and advanced analytics.

Which brings us to our second wave. We're going to continue improving this journey by embedding advanced analytics, which will enable our teams to prioritize cases more efficiently based on an enhanced risk classification. We also connect this journey to early warnings that will allow our analysts to act when a client has worsened their risk. We also adopt a swing line based approach, which means personalizing our journey so we can consider the proposal complexity, complexity and not only its value. We also adopt automated decisions, which will free up our analysts to focus on more complex cases. This path will also allow us to guide them through more efficient paths, while enforcing control and consistency for our third and last wave.

We're going to advance into GenAI Jill Power recommendations, decision, and Visual Assistant. All of that is going to be based on our recommended portfolio strategy, so we can have a healthier portfolio and also more sustainable in the long term. However, delivering the first wave, which, as I said, what we did right now was already instrumental. It allowed us to validate this journey that we constructed and also test user adoption and prove early impact.

So now let's take a moment to look at those changes. So guys, before we talk a little bit more about what you just saw, let's take a moment to think together. Let's pretend that our PegaWorld ends. You guys want to go to LA. What means of transportation. Are you going to use a car, a train, a plane, maybe a motorcycle? So that's the case here. The old system got us to where we needed to go because you can get there by various means. However, it was not the most efficient way to get there by time or experience.

So in the legacy experience, time was consumed more than necessary. It was fragmented and also error prone. In the new platform, analysts can now visualize the whole case with clarity, with information being centralized. And who better than to attest to that than their users themselves?

So let's talk a bit about what they are saying before launching Pega. We conducted a usability test because we wanted to confirm truly how intuitive the tool that we built was without training our documentation. 75 of our activities were concluded without any assistance or training, which was very encouraging for us. If you take the first feedback that we have here, that is our superintendent. He was the first one to approve a proposal on Pega and he was really happy about it. He could conclude it without any assistance or any training. And he had five people in the room with him, but he did it before anyone could explain it to him.

So as for the remaining 25%, we mapped the pain points to make the necessary adjustments in the journey and in the layout. We delivered that already in order to satisfy our customers. Now let's take a look at the metrics we have for our customer effort score. We reached a solid benchmark of 5.3. That meant our users enjoyed their experience and it was also easy to navigate. As you can see it, in the feedbacks that we have here, they call it simple, smooth and more efficient to work on and I swear we didn't bribe anyone. Those are real feedbacks from real users we have on the platform right now. And we are talking only about the first wave. So and we already got this kind of score right. Customer effort score. So very good. Perfect.

As for customer satisfaction score now. Well as you can see that improved drastically. We went from 1.75 with the legacy system to 4.25 on Pega. That showed how much better the experience has become and the journey has become for our users.

Now let's talk a little bit about the other results we have achieved. We have decreased 25% in our manual steps. We also reduced 75% in the systems used. That means fewer clicks for our users. Before, to conclude a single proposal, they had to use about ten tools. So they spend a lot of time, as you mentioned, collecting data instead of doing the analysis itself. So if this change, we allow them to focus where they should focus and not spending time preparing in order to achieve their goal.

We also had a reduction of 9% in rework, as we know we work, especially in credit analysis can mean frustration for teams and possibly delay for clients. So we had to work a lot on that and we are happy with the 9%. Where do we hope we will achieve more in the future? Because we structured data, the analyst can be sure he will receive all the information necessary to conclude the analysis, which was not the case before. And since we will move on to the other ways, we think that will increase even them. For.

Last but not least, as we know, time is a key metric for credit analysis. We reduce that on 20%. That means the closer we respond to our clients, the most likely we are to close deals. Perfect. And we are very anxious for the next waves. Because. Because we know that these results will be even better, even higher. Because we started we implemented the first wave on the smaller tickets, you know, clients with less complex credit structures. So on the next waves, where the credit analysts will have the support of the tool for like peer analysis, cash flow projections, we're going to plug in new functionalities. The gain will be much bigger. So it's we're very enthusiastic about it. We're very excited.

Yeah. And everybody's now going to talk a little bit what we learned during this process. Yeah. Every oops oops here. Every transformation let's say generates learning opportunities right. And uh, if I could give an advice for those implementing changes and, ah, the advices that we intend to use for the next waves is the first and most important one is, uh, conduct a thorough discovery process. I would say discovery. Discovery. Discovery. Don't assume. Use real people in the real process with the real pains they have in the process. Okay? And engage everyone in this process to ensure you deliver the right thing.

Because usually the. The business me like the business people, we don't like to sit down to discuss in detail. What we want. We just like I need a system to approve faster and like how you're going to do. And then ultimately the people from technology, from products, they they don't deliver what we need. Because we didn't describe well. So it needs to be perceived as a process that will benefit the. The organization that the client will be benefited by the by it and not like something that. Technology is going to deliver to us. Okay.

So um second MVP the MVP doesn't need to deliver. Everything at first, okay, you have to work in in waves, but the MVP needs to to provide an effective gain for for your customer, for your internal customer in this case. And they need to perceive the benefit out of any delivery OK effective communication like they need to understand everything you are delivering like this MVP will provide you this. The next one will get you what you ask for on on the other meeting. They need to know everybody that is involved, need to know what's going on and how we are going to to structure the deliveries, okay. And and communicate even what's not working. Right. We, we spend a lot of time adjusting the first release to ensure that we listen to everything that the analysts wanted. OK.

And, uh. But also equally important, robust partnership between business and it. Okay. It ties together with with what I said in the beginning. This is not, let's say, a solution for the business or for technology. It's for the bank. It's for the client. So everybody has to work together in partnership. Okay.

What else to wrap up? Um, what can I say? Transformation is you can deliver all at once. But when you have the right technology, you have the right people. You have a clear vision. Uh, I'm sure you can tackle the most complex processes. Okay, I think Bradesco took a very important step and supported by Pega. I'm sure that we we are going to evolve for a for a very innovative and very transformational digital phase in the credit, uh, in the credit team. Okay. Well thank you everybody.

Thank you guys. Thank you Mariana and Rubia for that. We said it's going to be an interactive session. So there's about 20 minutes for you to ask your questions. Please step up to the microphones and do who's daring to be the first one.

Q&A Session Yeah. See. Could you shortly state who you are when you want? Uh. Great presentation. Fantastic. I love the flow. The question I have is how long did it take for you to complete your way round after the discovery?

Six months. About six. Months. Six months. We. We delivered the first usability of the tool. And then how how how long now we have fine tuned to to go to for to roll out the rest. We're gonna roll out to everyone by the end of the year. OK seven will achieve the rollout of 70% of the volume by the third quarter. Right. And then for the entire segment by the end of the year. Thank you for that question.

Any more questions from your side? Yes. Please come up. Could you shortly state who you are or from which company you are to. Oh, I'm from Capgemini. Thank you. Yeah. So based on your experience on wave one, are you planning on expanding it to other use cases or other business lines?

Absolutely. I think it will be very easily escalated to retail banking and for the credit recovery process as well. So we'll be able to use a lot of what we did for the other segments and will be very easy to adjust. So I think we'll have a lot of gain from this process. So yes. Absolutely. Thank you. Thank you. Thank you.

Anybody else? More questions to come. We have time. Yes, please. I'm with Brian Advisory. What did you do with your data? Did you have to rationalize that across all the different systems platforms, or did you leave the data in existing silos? Sorry we didn't. Can you repeat your data? Yeah. So for the existing systems that you had there that you were transforming, did you have to move that data or restructure that data in any way to be able to use the new system.

We had to prepare it in order to connect it to Pega. We also had to transport it to Databricks in order to work with it.

Good afternoon. This is Steve Gross from Pega. So I have a two part question with respect to the accomplishments that you listed in terms of the percentage improvements in your next journeys, are you going to measure the same four areas to see if there's incremental improvement in those areas, or are you going to additional areas of measurement?

I think we'll have additional areas of improvement because we'll be able to incorporate new functions. OK. And the and the ultimate goal is to link not only the credit approval process, but also the the proposal part, the analysis part and the documentation and the completion part. So the end to end process. So I'm sure we'll have other things to measure. Got it.

And my second question follow on question was so it was good to have the percentage improvements. And you don't need to share with this large group. But have you also start to quantify the monetary improvements. Like what is the impact to the business in increasing the time to yes by 20%? You don't need to share it publicly, but are you putting dollars associated with those improvements as well?

Yes, we are measured by that because this is part of the transformation program, and we only manage to approve the investment and the changes because because we were able to prove that this would generate earnings, would generate a returns that would more than pay off the project. And it was one of the top projects prioritized because of its results. Oh, good. Well, I've asked ask that question. Thanks. Thank you for the question.

So do we have. Yeah. There's one more coming. Sorry. Yeah. Hi. Terry Holmes from Daimler Truck Financial Services. I was just wondering if during your discovery, if you used any of the Other Pega tools besides Blueprint, like any of the AI analysis tools, or what types of things did you use to analyze your current systems?

Yeah, so I forgot to mention the good question. We used the Blueprint any Other. But we focused more on the Blueprint. Yeah, at this stage for this first wave. Okay, thanks.

Hi. Justin Collery from Sunlife. So I see that you you had a green screen application, a mainframe application, and you migrated that. Yes. Did you retire the green screen application? Do you have an API to it, or is it running in parallel?

Not yet. We're in the process, so we need to roll out our new platform in order to retire the we call it black screen in Brazil. For obvious reasons. Exactly. And is there challenges in how you managing the challenges of running the two systems in parallel?

We have to do a constant monitoring of both tools. We have to ensure that the clients that are using Pega are no longer using the old system. But since we cannot close that so we can provide this monitor to be sure we don't extrapolate in any cases. So you've segregated. So we don't duplicate. Right. So it's it's it required some some adaptation. Okay. But the intention is to shut down the old system by the end of the year when we roll out the entire the Pega for the entire portfolio. Yeah. Thank you for the questions.

Any more from the audience. So I will have one for you. The easy one. The easy one? Yeah, okay. Fair enough. The easy one. First you talked about on the point three. The communication is so important and I've seen the results and they're amazing from the people. And you said you didn't do trainings. So what did you do?

I know we did trainings. Not just at first okay. The first thing we did was the usability test. But afterwards we trained everybody. We wanted to make sure they could use it all. So we did trainings, we communicated. We actually have constant meetings with them to make sure they know how to use it, and they are having a good experience. But but it's transformational. Imagine when you move, when you you get out of a blue or, sorry, a black screen and then move into a tool where you can, uh, include an organizational chart, you can attach a peer There are some industry comparables, whatever. You know, it's it's a totally different game.

So what are the big plans apart from wave three that you are going to incorporate in that storyline? We want to connect all the credit activities. So when they visit a client or when they respond to an email or all the documentation regarding the client, we want to build a one stop shop, as Ruben mentioned. So we want them to conclude the all the day to day activities in Pega. The spreading of the financial statements will be connected as well. And then cash flow projections and.

So did this question come up from here. Otherwise I do my last. I have a question. Yeah. Sorry. Go on. Congrats for the results and the presentations. Thank you. You said that you focus on Blueprint. How many tries or how many uses in the process you use because you use for the end to end process or for each process for each stage. How many tries do you do?

We use it for each stage, so we use it a lot and use it also to train our POS. So they use it in order to better understand the journey. So if I could guess like a 15 tries maybe. For each phase. For each phase, but just so they could actually learn how to use it better as well. Okay, thanks.

Thank you very much for your attention, for your time. Thank you. Thank you. Thank you guys. Thank you. Ruby and Mariana on that. And I want to remind you tomorrow at lunchtime, lending tables where you can meet, discuss, talk and so on. Have fun. Enjoy. PegaWorld. All right.