Success story

Banco Galicia onboards new corporate customers in minutes with intelligent NLP platform

 
 

Overview

Banco Galicia wanted to become the most customer-focused bank in its financial market and set itself a goal of reducing the time to verify new corporate clients from three weeks to just one. Working with Red Hat Consulting, it built an AI-based intelligent natural language processing (NLP) solution on Red Hat OpenShift Container Platform, Red Hat Integration, and Red Hat Single Sign-on (SSO). The bank has not only cut verification times from days to minutes with 90% accuracy, but it also cut application downtime by 40% and improved agility fourfold. The NLP platform also represents a significant step in the bank’s journey to transform its technologies, processes, and culture.

2023 Innovation Awards Winner


 

Challenge

Onboarding corporate clients slower than the competition

Banco Galicia is one of Argentina’s leading private banks, with 5,400 employees serving 3 million personal banking customers and 150,000 businesses of all sizes and across all sectors. Its services span everything from financing to billing and payments and even investment. It also has foreign trade products for customers who operate abroad. A traditional bank, Banco Galicia is a leader in technology innovation.

Banco Galicia’s Official Acceptance of Credentials (OAC) process determines whether a corporate client can perform financial transactions with the bank. The process, part of the critical ‘Know Your Customer’ (KYC) model, took up to 20 days to complete – considerably slower than the competition.The legal, financial, and economic documentation was unstructured and contained a lot of human language; it had to be processed, analyzed, and verified manually by different teams. 

The bank wanted to become the most responsive bank in its financial market and the fastest to verify corporate clients. As a first step toward what is known as ‘Bank 4.0’, the bank set itself a goal of reducing verification time to one week. A new project sought to reduce the number of days for OAC processes so new corporate customers could access the bank’s services faster.

Solution

Building an open source intelligent NLP platform

Banco Galicia worked closely with Red Hat Consulting to build an AI-based intelligent NLP platform. Red Hat OpenShift provides the on-premise platform with a reliable and highly available foundation, with the AI workloads deployed in containers to optimize inference processing.

To develop and train the AI models, Banco Galicia’s data scientists use a comprehensive set of open source AI tools from the Open Data Hub (ODH) open source project. This includes Jupyter Notebooks (for development) and  TensorFlow and scikit-learn (for training), which all run on Red Hat OpenShift. 

Continuous re-training the API-based models allows the bank to improve processes, and then make the models available for other projects and areas. 

Banco Galicia also takes advantage of Red Hat Integration and Red Hat Single Sign-on (SSO) to further streamline the platform. Red Hat Integration technologies expose models to any internal workloads not running on the cloud, while Red Hat SSO simplifies and secures access across the Red Hat technologies.

Dive deeper into Banco Galica’s NLP solution

Hear more from Banco Galicia on how they built their AI/ML solution and what the future holds for this innovative project.

Business outcome

Cut onboarding times from 20 days to minutes

An initial proof-of-concept reduced the data processing time for verifying corporate clients from days to minutes. It also achieved 90% accuracy, exceeding the project’s initial goal of 80% and proving that reducing analysis times was feasible.

Now, waiting 20 days for verification is a thing of the past. The bank’s corporate customers enjoy an onboarding process that is 100% digital, transparent, and self-managed. Analyzing documents is practically a real-time process  where customers can open and start using their account in minutes. 

Likewise, employees previously dedicated to analyzing and processing documentation manually have been assigned to more analytical tasks, improving job satisfaction. By reducing OAC process times, Banco Galicia  has positioned itself as one of the most agile banks in corporate banking.  In fact, automating the process has helped the bank to save around 40% on operating costs.

Built with Red Hat technologies at its core, the intelligent NLP platform gives Banco Galicia the freedom and potential for growth; it owns the knowledge, the algorithms, and the data for training algorithms. The project represents  a significant first step in the bank’s digital transformation journey, initiating a shift in its technology, culture, and processes. This progress is also critical to helping the bank maintain a leading position in the financial industry and technological innovation, as well as get closer to achieving its Bank 4.0 ambition.

Icon-Red_Hat-Media_and_documents-Quotemark_Open-B-Red-RGB Red Hat plays a role in implementing NLP within Banco Galicia by providing us with the technology and architecture. Through Red Hat, we managed to understand everything to do with Red Hat OpenShift. We also started to design an architecture that could be cloud-native.

Matias Lorusso

Solution architect - Banco Galicia


Related resources

e-book

Top considerations for building a production-ready AI/ML environment

Red Hat provides a complete technology portfolio to help you achieve your AI/ML goals.

E-book

Build data-driven applications with containers and Kubernetes

Gain a competitive advantage with databases and analytics.

details