 We're back, Paul Gillan with Rob Trichet here at the Red Hat Summit and Ansible Fest, 2023 in Boston. And we've been talking a lot about digital analytics, about how data is transforming organizations. And we have a company here that's walking the walk. Joining us right now, Banco Galicia, out of, did I pronounce that correctly? Galicia. Galicia, out of Argentina. Ericko Bemmer is the Chief Information Officer. Victoria Martinez is the Latin American Manager for Business Development for Data Science at Red Hat. And the bank has recently completed, you've been a Red Hat customer for a long time. I see several case studies on Red Hat's site about the bank. You recently completed a digital onboarding initiative. Tell us what the objective, what was the goal of that initiative and why did you make the investment at this point? Okay, well, Banco Galicia is an 100% Argentine on coveted banks. Back in 2019, we were facing an ambitious to grow our market shares in the corporate sector. So we had a really manual process in the onboarding because there is a lot of information and papers and documents that we have to process manually to onboard a company, you know, statues, powers attorneys, these kinds of documents that are all papers, paperwork. So the whole process to onboard a company, it was more or less to 20 days average. So it was really tough for us. It was an expensive process. And of course we used to lost a lot of opportunities in companies that abandoned the process in the middle because of the time. So we realized that we had to work deeply in a new solution in order to shorten that time and to improve our customer experience. How did you really look at, did you look at, I guess, did you really look at the whole process? And I think, again, you brought in Red Hat and as Paul was saying, you have a long relationship. How did you work together to really get through this? Because when I went and looked at the little case study, there's a lot of technology there, not just OpenShift for data science, you know, there's a lot of, and I think there, what will be interesting to the folks watching is they're just starting their AI journey. And I think to your point, you have a really good use case, I think that you started with kind of that use case. What was it that Red Hat helped you with in getting started and getting going with that? Well, to be honest, this project began a day before that and we were trying with different technologies, more closed, more black boxy technologies. So we had a couple of attempts failing and we were really frustrated because we wanted, and we believed that using NLP technologies, we could reach the digitalization so in that time, the Red Hat consultant team came up with a solution to change the whole stack that we were using and to start building the project over OpenShift. We were using OpenShift for a long time. We had almost 25 clusters right now. We have 300 projects, 15,000 bots running every time. So we were really excited to change the technology stack and to start building with a different mindset, with an OpenShift mindset, so that's why, how we started. Victoria, describe the process by which you become involved with a project like this. Does the bank come to you and say, this is what we want to do and how will you help us? Or where were they in their journey when you became involved? Yes, also. We start with these two fail attempts, particularly NID with the digitalization, with NLP. We share this vision with our world, with the open source technology, bringing these workloads with machine learning, the models and the data stuff on the containers. This was a new way to concept this vision with the AI technology because when we start, was a different reality. Today we have the transit topic with the GDP, with the AI, we are living in the boom, but that this those years ago, we share a new vision and a different thing with the project. We want to build the capability with NLP in the bank, not only a project, not only solve a use case, was totally different with another experience because there was close, the thing with the compliance and the customers, very sensitive with the data, all the solution is on-prem. We share this new vision, start with a little use case and with a little MVP, and then we large to all the onboarding process. So you built the NLP from scratch? Yes. What role does NLP play now in the onboarding process? Well, it's one of the central pieces of the whole process because to analyze the whole document and paperwork, today needs different technologies like AI. So the process is more or less in four parts, steps, the first steps is the processing of the images with computer vision, the libraries and plugins, in order to clean the images, I don't know, delete the stamps, rotates, change contrast of the documents, in order to the second step, which is the OCR, to capture better each letter. Then is the magic of the NLP, trying to put the letters together to form words and using, of course, natural language processing and to complete words with letters that we missed during the OCR process. And at the end is the entity recognition and mapping that close the whole magic process in order to capture name of the company, name of the powers of journeys, I don't know, the activity of the company, balance sheets, process all the whole balance sheets, healthy situation of companies. And those kind of information is all unstructured in just images, PDF files. So this is the important thing regarding to AI, which is not simply to use another technology to do that. It seems like you started right, sounds like right before the pandemic hit. And it would seem that this became probably more, even more important during that, because then you would have less people involved in the entire process. Did that help accelerate the investment that the bank made into this project? I actually, as you told, the project began before pandemic. And during the pandemic was one of the key pieces to maintain the growing and growing by the time. Because by the time we were releasing the first step of the projects and automating the whole process end to end. And yeah, it was definitely a key factor in order to keep our growing and to maintain the customer satisfaction and the satisfaction of our customers in the pandemic was really tough for people and for companies too. You chose to use a number of cloud native constructs in building this using APIs, containers, OpenShift, Kubernetes. What benefits do you expect to realize over the longer term from using these cloud native constructs? Well, in one hand, you know, to build with calculated technologies in containers. Nowadays it's running on-prem, but for example, one of the travels we faced during the project was we couldn't reach the computational power to train the models. So it was easy to move those components to the cloud, use power consumption, and then go back to our own prem in order to apply the models already trained. So I think that this is the facility and the benefits of clouds. And in the other hand, we build a platform API oriented. So by the time, nowadays, GPD and other technologies that we can plug in and renovate, you know, modernized technology using different kinds of components, new ones to try different NLP models, of course, GPD to recognize new entities or words. So it's like a Lego system. It seems that flexibility and the ability to burst up into the cloud when you needed to do the big training portions of it also would help you be more efficient in how you use your resources on-premise. But is that something that you look to continue to do? Is that, so as you, like you said, build another part to the Lego, add another Lego block in, are you using still doing that bursting to certain, to cloud to go and do that? Well, we have, of course, our journey to the cloud. As many, many companies. As we are a bank, we are a regulated company. So we are trying to move consistently, but security. So we are adopting different parts of the clouds by the time. At the same time, we are rebuilding and architecture our channels, for example. So yes, we are expanding the possibilities to move these technologies and to keep growing and growing, scaling to the public cloud. Victoria, as someone who works with, who knows the Latin American market very well, can you contrast what is the state of open source in Latin America compared to North America? Is it any different? No, the innovation is in the communities and in the open source. This globalization that we live, don't have a gap between us, between Latin America or North America or the Europeans. We live a new way to concept the technology and to interact with the users, with the velocity. It's the same for every country, for the complete world, for the global thing. It's the same need. We are living in a new era and we need to accelerate this thing and improve with open source. We in Red Hat, trying to accelerate the use cases with machine learning ops, with OpenShift data science and the new with OpenShift AI, trying to connect the new technology stock for the delivery and application and intelligent application with models and machine learning and AI. You heard about OpenShift AI at the summit this week. What's your reaction to it, Eriko? Is it something that you would plan to adopt quickly or perhaps you already have? Well, actually, we begin building this project before the OpenShift AI exits, but we are really excited to try new technologies to go further, you know? So, to these days, there was a lot of launches, like AI, Ansible Automation with Event Driven. So, we are really excited to try all those technologies to improve the modernization of the company. It seems like this is a, you're actually well ahead of even some of the ones that I've talked to in the US where they're still trying to figure out their cloud journey in the financial services. I talked to a lot of even some of the largest banks in the US, haven't even gone to certain, like haven't moved to cloud databases or anything of that nature. Do you feel that you're way out there on the edge or do you feel like you have this all under control because of the help you've gotten out of Red Hat? Actually, we say that we have our own basing, so we are moving firm and security, but we are not facing that it's out of control. We are controlling the technology and not the technology controlling us. That's great. Well, it's a terrific story and if you want to read more about it, it's on the Red Hat side, Bank of Galicia Streamlines digital onboarding with AI platform built on Red Hat technologies. There's quite a bit of detail there about what they're doing and you are out on the leading edge whether you think of it or not. Thank you so much for sharing your story with us. Well, thank you very much, Paul and Rob, for the interview. Paul Gillan and Rob Stochet at Red Hat Summit 2023. Wrapping up day two, we'll be back in a minute.