BCI transforms core systems to unlock credit card transactions

THEIR CHALLENGE

Breaching BCI's banking bottleneck

Founded in 1937, Banco de Crédito e Inversiones (BCI) is one of Chile's most prominent private financial institutions. With expanding global operations across its home country, Peru, and with the acquisition of City National Bank of Florida and BCI Miami in the United States, the bank strives for optimal customer experience and operational efficiency based on data and AI. 

However, an impressive 18% annual growth rate in transactions and clients within the bank’s wholesale division placed an immense strain on its legacy core architecture. BCI's critical system responsible for serving credit card and debit card statements was hitting technical limits under an on-premises HBase setup. 

“We were experiencing performance issues when our larger customers tried to query historic data,” explained Alvaro Olivares, Technical Architecture Lead at BCI. “The response times were outside of what they wanted, errors were happening – it was a pain.” 

As a Tier 1 system, BCI's statement platform must adhere to strict regulations from Chile's Comisión para el Mercado Financiero (CMF), including proactive reporting of any unexpected system downtime. Crucially, such structural performance bottlenecks meant BCI could only keep a restrictive six months of transactional history available online for real-time customer queries. In addition, the heavy infrastructure costs and operational complexity of ‘firefighting’ the rigid HBase environment made it economically impossible for the bank to efficiently scale its digital banking services to small and medium enterprises (SMEs). 

"Our current rate of growth is a big challenge because all the core systems that we have are very old and we are very close to their limits," said Andrés Mella, Enterprise Banking Lead at BCI. "All these projects are needed to keep that rate of growth and make it possible to maintain in the short and medium term." 

OUR SOLUTION

Migrating to a resilient, multi-region architecture

To break free from architectural limits, BCI migrated its statements transactions platform to MongoDB Atlas on Microsoft Azure. The bank's engineering teams collaborated closely with MongoDB experts to finalize optimal platform sizing, design deployments, and implement best practices. To guarantee a seamless migration without disruption to live banking operations, they ran a parallel dual-write replication strategy to sync both the legacy HBase database and the new cloud environment until historical and real-time data sets were fully validated. 

Once live, MongoDB Atlas functioned as a high-performance data access layer that insulated legacy core banking systems. By shifting high-volume read traffic to MongoDB Atlas, the core accounting engines could focus solely on transaction processing. 

“When we did that, response times were faster, and we could scale better when we had peak traffic at the end of the month or when closing the fiscal year,” explained Olivares. 

To deliver this horizontal elasticity, BCI used MongoDB Atlas's fully managed sharding, giving the team the ability to dynamically adjust capacity ahead of intense commercial spikes like Black Friday. They also leveraged secondary indexes to accelerate specific, heavy filtering queries that previously triggered system latencies. 

Finally, to satisfy CMF compliance and guarantee high availability, BCI established a multi-region cluster deployment spanning Azure Virginia and Azure Santiago. 

“This is something we deployed after we migrated everything to MongoDB,” said Olivares. “The capability is soon going live where we’ll be serving our customers in Chile from the Chile region. This will further improve the response time and give us better resiliency.”

OUTCOME

Faster statements, zero timeouts, and 130,000 new clients

By transitioning to MongoDB Atlas, BCI has created significant operational efficiencies, transforming its core systems and product features while accelerating overall business growth. Most significantly, MongoDB Atlas's cost-effective scaling has enabled BCI to expand its target market beyond an initial 20,000 large corporate accounts, paving the way to easily onboard an additional 130,000 SMEs that were previously too expensive to serve on the old legacy setup. "To reach this segment of companies [SMEs] is a very big achievement of this project,” said Mella. “That wasn’t possible before MongoDB.” 

By moving to a fully managed platform, BCI has lowered its total cost of ownership (TCO) by 25% and reduced day-to-day maintenance complexity. This streamlined operation has directly supported BCI's efficiency targets: by enabling corporate clients to completely self-serve statement queries on the 360 Connect enterprise web portal, BCI successfully contained the headcount growth of its operational support teams. 

MongoDB now handles an incredible 10,000,000 transactions per day, absorbing traffic every time a retail or wholesale banking customer logs into the mobile application or business web portal. Even under this immense operational load, system response times have accelerated dramatically, operating two to three times faster than the legacy HBase infrastructure. Heavy data batch processing times have been cut by 40%, from 50 minutes to just 30 minutes, while long-tail timeouts for heavy municipality and corporate users are at practically zero. Freed from past performance limits, BCI has confidently extended its online real-time data history window from six months to 18 months, fulfilling the intensive visibility demands of its largest wholesale clients. 

As BCI looks to the future, the absolute data accuracy provided by MongoDB Atlas gives the bank the confidence it needs to safely explore new open banking models. And its robust cloud architecture provides a resilient foundation as BCI next prepares to handle an influx of sophisticated agentic AI queries across its multi-region nodes and systematically migrates its massive core batch payment processing pipelines onto MongoDB. 

"The stability and certainty that MongoDB brings to the table gives us peace of mind,” concluded Olivares. “We’ve seen good results, better customer fit, and lower operating complexity and cost.”