Teradata Possible 2024 | Michael Hutchinson, Teradata & Matheus Daniel Pierozan, Sicredi

Clips
News
More from Teradata Possible 2024

Michael Hutchinson

Chief Customer Officer

Teradata

Matheus Daniel Pierozan

Senior Credit Analyst

Sicredi

Banking world sees faster AI deployment with Teradata

Teradata Corp.’s focus on hybrid data analytics has extended to the banking world.Sicredi Confederation, a Brazilian credit union and one of the top ten banks in Brazil, is working with Teradata in association with Amazon Web Services to improve its customer experience. Teradata’s software has led to rapid improvements in Sicrediti’s data management, according to its analysts.Teradata’s Michael Hutchinson and Matheus Daniel Pierozan talk with theCUBE about how Teradata has changed the bank’s operations

play_circle_outline Introduction of Sicredi, a Brazilian credit union, and their use of Teradata

play_circle_outline Benefits of using Teradata for data processing and modeling techniques

play_circle_outline Results seen by Sicredi in terms of client income prediction and credit limits

play_circle_outline Future goals include focusing on modeling techniques and showcasing more use cases

Info
Transcript

Michael Hutchinson, Teradata & Matheus Daniel Pierozan, Sicredi

Michael Hutchinson

Chief Customer Officer Teradata

Matheus Daniel Pierozan

Senior Credit Analyst Sicredi

The event included Mateus and Mike from Sicredi and Teradata. Sicredi, a Brazilian credit union with 8 million clients, has a 20% annual increase and is one of the top 10 banks in Brazil. They chose to work with Teradata for speed and best practices application, focusing on real-time information for activities like credit approvals. Accurate data is crucial for financial solutions, with improvements seen in consumption limits and processing time. Mateus aims to enhance modeling techniques, while Mike focuses on showcasing successful use cases with ClearScape ... Read more

explore Keep Exploring

What led the organization to choose Teradata as a data processing solution, specifically in association with AWS? add

What are some of the challenges and solutions that are frequently encountered when working with financial institutions like Sicredi in terms of data processing and speed with Teradata? add

What impact did using Teradata have on increasing client income in the first month? add

What is the goal of running AI innovation days with customers in relation to ClearScape Analytics? add

bolt Powered by CUBE AI

Michael Hutchinson, Teradata & Matheus Daniel Pierozan, Sicredi

search

Savannah Peterson

>> Good afternoon Teradata community and welcome back to Los Angeles, California. We are here with our final segment for what has been an extraordinarily educational and inspiring day. My name is Savannah Peterson, joined all day with Rob Strache. I can't believe it's coming to an end.

>> I can't either, because I love talking about data. I love talking about the use cases and how customers are really just diving in and how they're being successful with it. So this is fantastic that this is the way we're ending today.

Savannah Peterson

>> I know, it's like a little cherry on top. Mateus and Mike, thank you so much for taking the time to join us. Busy day for you both, I bet.

>> It's been a bit busy, no doubt. But glad to be here.

Savannah Peterson

>> Yeah, I can imagine. So I'm going to start with you. You're our first Brazilian guest today, and I'm excited because Teradata touches customers all over the world. You've got employees in six continents. I mean, yeah, exactly, I said that right. And so just an example of the global reach. Tell us a bit about Sicredi before we get into the beat.

>> Yeah, Sicredi is a credit union from Brazil. So we have operations in all the Brazilian states, our focus in rural lending credit, and we have now 8 million clients.

Savannah Peterson

>> Wow.

>> Yes, we are increasing around 20% a year in terms of clients.

Savannah Peterson

>> And one of the top 10 in Brazil as well.

>> Yes, we are in the top 10 banks in Brazil.

>> So what brought you to Teradata? And I know you're doing it with an association with AWS, how did that really come about and why Teradata?

>> Yeah, I think in a data scientist's perspective, I think we are using Teradata and the power of the data processing, especially because of the speed, that this is the biggest advantage of using Teradata and we are able to apply the best practices with a low learning curve.

Savannah Peterson

>> Speed has got to be mission-critical. I mean, talking about banking, nobody wants anything slow. If I want my money or if I want to get approved for a credit card, whatever that might be, I want it yesterday. Mike, I'm sure you see a lot of trends with about 50% of Teradata's customers being in the financial institutions. What are some of the challenges or solutions that you're architecting most frequently with these types of customers like Sicredi?

>> So a lot of the work we do is in the commercial banking space. So a lot of large banks, they are serving customers all the time. And as you just said, customers are looking for real-time information around the activities they're doing in the bank, and if they've applied for credit at the bank, they're looking for a real-time decision. Am I going to get the credit? When am I going to get credit? And more frankly, when am I going to get the money? So we're doing a lot of work with a lot of the financial institutions globally on a whole set of use cases. And many of them will be around the customer experience. It's a big area for us. But just equally, there's a number of use cases around risk and fraud. You're in a large economy somewhere like Brazil or anywhere in the world, the banks have an obligation to protect their assets as well. So the use cases span the gamut in terms of what we see as we work with the banks like Sicredi.

Savannah Peterson

>> I believe that.

>> And with Sicredi, what is it if you don't have accurate data, because I think a lot of what we've been talking about is really accurate income estimation, things like that. What does it mean? Is it higher interest rates, what are some of the examples of not having the right data?

>> Yeah, the income information is the main information to offer financial solution to our clients. So if you have a bad prediction of the income information, you will probably offer a product or a service that won't fit in the client's needs. This is why it's so important to have good predictions.

Savannah Peterson

>> So when you started, you're growing quickly, looks like, I mean, I see in here 13% annual new customer growth, not even just building up your existing customers. That's pretty massive scale. What have been some of the results that you've seen since you started working with Teradata to evaluate someone's credit profile?

>> In this case that we are trying to predict the client's income. Just in the first month, we were able to increase the limits consumption in 30% using Teradata. Yes.

Savannah Peterson

>> In the first month?

>> Yeah, just in the first month. When we deployed them-

Savannah Peterson

>> 30% in 30 days. I bet everybody wants to achieve results like that.

Savannah Peterson

>> Yeah, we've done a great job in terms of modeling improvements as well, and also using updated databases, retraining the model. So we work on that and we also improve the number of clients that are being scored by this model. We've done a great job.

Savannah Peterson

>> So this bring your own model announcement that Teradata had today, you were already working on it. What does it mean for you, Mike, to be able to bring that to more customers?

Savannah Peterson

>> So I mean it really follows Teradata's strategy overall is look, we know there are things that we do really well and our job is to provide a platform for customers come to that is going to be able to handle massive amounts of data, is going to be highly scalable, highly performant. But what we also know is we want to be open and connected. So if you have a model that is working and you bring it into Teradata, and that's exactly the path they followed, what did they get out of it? They were able to get a good model, but the performance that they saw in running the model was fantastic. And I'll let him share that with you. But our goal is to make sure that we've got a platform that's open, allows you to bring your own models connected, allows you to connect with other third-party services, to really drive business value like they're being able to do. I think the one thing that I just got to see his presentation earlier this afternoon is not only are they doing a much better job serving the customers, but when you look at what they've done, they've also brought down the risk profile of the bank, which is just hugely important. So serving more people at the same time, normally you're going to think, "Hey, I'm serving more people, I'm increasing the risk to the institution." They've been able to double down, serving more people and actually bringing down the risk, and very quickly.

Savannah Peterson

>> It's optimizing both ends of the business really.

Savannah Peterson

>> It really is. They've done a great job.

>> So what was it like before you got to ClearScape Analytics bringing your own model?

Savannah Peterson

>> Yeah, the before scenario was hard because we had an uncalibrated model, codes with low performance. We had a no control on data transfer and a limited target group of clients that were being scored by this model. So it was necessary to rethink this process.

Savannah Peterson

>> Yeah, I can imagine. How long ago did you end up working together? When did this engagement begin?

Savannah Peterson

>> In this case?

Savannah Peterson

>> Yeah.

>> I think almost four months.

Savannah Peterson

>> That's it?

>> Yeah.

Savannah Peterson

>> You've already seen these results. This is insane. I'm looking at some of the numbers in my notes here. I mean, I am not even trying to over exaggerate, but this, I mean just-

>> Four months.

Savannah Peterson

>> Just speaking about models. I know that Teradata is size-inclusive, and this BYOM bit is really important, these are massive informants. 40% decrease in time spent to build, test and deploy, leading to 25 X faster model processing.

>> Yeah.

Savannah Peterson

>> 25 X.

>> Yes. Yeah. Our former model, the time of processing the entire model from the beginning to the predictions, it was around two hours and 10, two hours and 30 minutes of model processing. And now with Teradata, our best time was just six minutes.

>> Wow. And was that preparing the model, training the model, all of that? Or is that-

Savannah Peterson

>> Yeah, just model processing. Yeah.

>> Oh, just model processing.

>> Yeah. From the beginning, querying the data, transforming, preparing the data to enter in the model processing and then getting the predictions.

>> Wow.

>> This is the entire path of the data and this is the outcome.

>> And it wasn't like-

Savannah Peterson

>> That's impressive....

>> before you were using some arcane technology or something like that, you were using Databricks and you were using a cloud-based model. What was the big change, I guess you could say when you moved from there and how long? Because I mean you were building the models and to your point earlier about data science, you have to try retrain, go back, refactor the models and do all of the model pipeline. What really changed with it?

Savannah Peterson

>> We have the custom of always trying to test all the technologies that we have. And we do that in small queries, small tests enough just to test. And it's always in, can I say this? It's always small tests and a fair comparison to have the results and compare, "No, this is better, this is not." So this is why we've chosen to process everything inside ClearScape Analytics.

Savannah Peterson

>> Makes a lot of sense. What problem are you going to solve next using Teradata?

Savannah Peterson

>> Yeah, we have a challenge, a lot things to do next year. We have several new regulatory models, and we are trying to improve the quality of the data that will be used as an input. So we are preparing this book of variables. It'll be challenging because now we have 900 variables. So we are trying to increase the number with quality and speed.

Savannah Peterson

>> Just a couple, just 900 different, choose your own adventure little pathways that could fork in, yeah, no big deal, just a small amount of complexity.

Savannah Peterson

>> That is a lot job to do.

>> How does this help customers in that way? I mean, again, the time to value seems to be a very large time to value.

Savannah Peterson

>> I mean, that's the obvious one, it's time to value. What does that mean? It means they can run their models much more often. They're freeing up resources on the systems for other use cases, but there's also some other benefits that it's just like, "Hey, we got to more customers and we lowered the risk." There's other benefits that come through this as well. So one, they're saving money on data transfer charges. Previously in their data model, they were moving data back and forth. That's a cost issue. And guess what? It's also a security issue. By keeping the data all with advantage. One, you're avoiding the cost, but two, you're making it a much more secure situation. And as a bank, we've talked about it, security is paramount above everything else. So you always go to the end goal and like, "Oh, it's better timing, better performance," but you step back, your security guy is happy because, "Hey, look, we're doing lesser." Your CFO is definitely happy because we're spending a whole lot less money to get this. Your revenue team is happy because they're getting more customers, more revenue. So there's so many good angles to this. And what I always encourage customers is pick what you're driving for, but don't forget that simplification can have a lot of big benefits across the ecosystem that you may not even be part of your ROI.

Savannah Peterson

>> I think that's a really... I mean you just brought up a really good point, , it's look at the ripple effect of what good can happen. It can become a more sustainable ecosystem. You're spending less money, it's taking less time to run models. I mean, you win in every aspects.

>> Yes. Yeah.

Savannah Peterson

>> Which is, wow, . All right, I have one last question for you because you're both great. You're doing awesome. I know it's your first time up here and I just got to give you a shout. You're totally doing great. What do you hope to be able to say next time we're at Teradata Possible having another chat? You've had this four months of adoption in this particular use case. What do you want to be able to say a year from now that you can't say today?

Savannah Peterson

>> I think in my perspective as a data scientist, I think I would love to say that for us, in my position, it's important to focus on the modeling technique. We are concerned to develop a better model, to improve the model, not just be concerned in terms of the data pipeline, because this takes a lot of time of work, 90% of the time, 80% of the time, treating the data. And sometimes we don't have time to focus on the modeling technique, different approaches. And for me, this is something that is really important that I would love to say.

Savannah Peterson

>> Great. Well, we hope to be able to have that conversation. What about you, Mike? What do you hope to be able to say?

>> So look, I just want to be able to say, "Here's 10 more, 100 more examples of great use cases that have come out by employing ClearScape Analytics." The way we are really trying to get to those is our customer success team, they're out with customers every day running AI innovation days. And it's things like these that are coming out of the innovation days to say, "Hey, how about trying this? Have you thought about using ClearScape for this?" Or, "How about a tweak to this model?" Or, "Hey, here's a brand new use case, and you've got most of the data already in the system." So our goal is running the innovation days, is to have more and more and more to build on examples like this because they're truly great stories of the capabilities of vantage and how vantage is delivering real business value and outcomes to customers like Sicredi.

Savannah Peterson

>> Well, we look forward to telling all of those stories. Mike, Mateus, thank you so much for being here. This was awesome. Rob, it's been a really rad day. Also, shout out to the production team over there, making sure we all look fantastic. And everybody else back at the ranch at theCUBE. And most importantly, thank all of you for tuning in to our fantastic day of power-packed coverage. We've done nine interviews here today at Teradata Possible. I feel smarter. I certainly hope you do. We're in Los Angeles, California. My name is Savannah Peterson. You're watching theCUBE, the leading source for enterprise tech news.