- Feature lead time reduced from 90 days to 30 days.
- Historical data access expanded from 3 months to over 5 years.
- Response times slashed from 2 seconds to 200 milliseconds.
- Mainframe MIPS consumption reduced by over 9%.
Itaú, one of Latin America's largest private banks, has embarked on a remarkable journey to modernize its core banking services, specifically focusing on bank statements. Faced with the limitations of legacy, monolithic architecture, the bank sought to transform its customer experience by embracing cloud-native solutions and artificial intelligence. This ambitious undertaking has not only yielded significant performance improvements and cost reductions but has also paved the way for innovative features like intelligent search.
Before modernization, Itaú's bank statement service suffered from high latency, limited historical data access (only three months online), and an inability to perform complex searches or data aggregations, particularly for corporate accounts. This resulted in degraded performance during peak demand and a less-than-ideal customer experience. Recognizing the need for speed, personalization, and seamless interactions, Itaú defined a clear vision: a new unified platform serving all channels for both individual and corporate accounts.
The new architecture, built on AWS, prioritizes five key technical drivers: resilience, scalability, cost optimization, security, and the ability to enable new features. A critical component of this transformation is the use of Amazon OpenSearch Service as the data layer, replacing traditional relational or NoSQL databases. To ensure high availability and reduced impact in case of failures, Itaú implemented advanced resilience patterns like shuffle sharding and hedging, where customer data is replicated across multiple shards and requests are sent to multiple shards simultaneously, picking the fastest response.
The outcomes of this modernization are impressive. Feature lead time for new functionalities has been drastically cut from 90 days to just 30 days, demonstrating a significant increase in agility. Customers can now access over five years of historical data online, a substantial improvement from the previous three-month limit. Performance metrics show a leap from 2 million requests per second monthly to over 8 million, with response times plummeting from around 2 seconds to a mere 200 milliseconds, even during high-demand events like Black Friday. Furthermore, the migration has led to a reduction of over 9% in expensive mainframe MIPS consumption, highlighting substantial cost savings. The improved transaction descriptions and near real-time corporate account reconciliation APIs have also boosted customer satisfaction and engagement.
Looking ahead, Itaú is rolling out its AI-powered intelligent search, utilizing Amazon Bedrock to process natural language queries and translate them into optimized OpenSearch queries. This allows customers to ask questions like "how many Pixes did I do last month?" and receive instant, accurate results, reducing errors by 80%. The ultimate goal is to build a conversational bank, where natural dialogue and device commands become the new normal for customer interactions, further solidifying Itaú's position at the forefront of financial innovation.
“Modernization is a journey. This is not one-time event. You need to keep evolving the platform every day. You need to optimize performance. You need to improve the scalability. You need to strengthen the security. You need to create new features for the business and so on.”
- Eduardo Rodriguez, Principal Solutions Architect at AWS
