Empowering Personalized Banking Experiences

At Zafin, our mission is to help banks modernize their core infrastructure to deliver exceptional, personalized experiences to their customers. To determine customer relationship tiers and provide personalized rewards, we needed to process significant data volumes—100 million accounts with over 10 billion balances. Traditionally, we would have implemented this tiering application using an open-source framework, Java Spring Boot, with a dedicated PostgreSQL server. However, building and scaling this solution with a traditional tech stack proved inefficient and unsustainable at scale.

Traditional data infrastructure doesn’t scale

Like many organizations in financial services, we initially relied on dedicated data warehousing solutions to power our analytics. While functional, this approach came with significant challenges: high compute costs, maintenance overhead, and performance bottlenecks.

With a single-tenant architecture, every customer deployment requires dedicated infrastructure. Since some customers use analytics sporadically while others rely on it actively, this setup was expensive and inefficient. Dedicated infrastructure demanded constant fine-tuning and incurred fixed costs, regardless of usage—a significant effort for our engineering teams. Scaling to accommodate massive data volumes across millions of accounts, balances, and transactions often stretched our traditional systems to their limits.

Complex computations, such as processing billions of records for customer segmentation, tiering, and analyzing relationship patterns, ran as batch jobs that were slow to complete. These bottlenecks delayed time-to-insight, and our traditional approach couldn’t keep up with the scale and speed banks required.

To seamlessly provide advanced analytics to our customers, we needed a cost-effective, scalable platform capable of handling massive data volumes, delivering high performance, and keeping costs under control.

Migrating to Databricks to accelerate data intelligence

We ran a proof of concept (POC) to compare our previous Java Spring Boot/PostgreSQL setup with the Databricks Data Intelligence Platform. Here’s how the POC empowered us to improve our data infrastructure.

Databricks delivered up to 10x faster computation speeds compared to our previous solution. For example, tasks that previously took 4,000 seconds on our previous system are now completed in just 300 seconds. With elastic scalability, we can process billions of transactions and balances across millions of accounts without compromising performance.

Migrating analytics workflows reduced time-to-market for new capabilities by 30–40% while requiring fewer engineering resources. Compared to our traditional setup, Databricks helped us accelerate the delivery timeline of our tiering application by 30%. With less effort spent on infrastructure management, our engineering team can focus on building innovative solutions for our customers.

Databricks’ Serverless capabilities were a game changer, allowing us to separate compute from storage to ensure we only paid for the compute we needed. This approach eliminated the cost burden of idle infrastructure. Cloud costs for analytics dropped by 50–70%, depending on usage profiles. The ability to auto-scale based on data loads meant that both high-traffic and sporadic-use customers could rely on a seamless, responsive experience without manual tuning.

The results were clear: Databricks offered a transformative solution that delivered superior performance, scalability, and cost-efficiency—all while meeting the stringent governance requirements of highly regulated financial institutions.

Delivering value to our customers

Databricks’ unified environment allows our teams to handle data ingestion, transformation, and analytics in one place—improving productivity and collaboration. From data ingestion to complex SQL-based transformations and embedded analytics, everything runs seamlessly on Databricks where even non-technical users can derive meaningful insights. When speaking with Terry Hickey, Zafin’s Chief Revenue Officer, he expressed, “With the modernization of our data infrastructure, we can now help banks effortlessly uncover actionable insights to drive growth and deepen customer relationships.”

With Databricks as the backbone of our modern data stack, we’ve been able to innovate faster and deliver greater value to our customers, including:

• Personalization: Using Databricks SQL and Notebooks, we built a data-intensive tiering application that evaluates customer balances, product holdings, and transaction patterns to determine relationship tiers (e.g., Gold, Silver, Platinum). This allows banks to deliver personalized rewards and benefits to customers while nudging customers toward the next tier with targeted engagement, increasing retention and loyalty.
• Business Intelligence (BI): Our embedded analytics solution, Zafin Analytics, helps line-of-business bankers gain actionable insights into product performance and offer effectiveness. With Databricks’ integrated environment, we provide out-of-the-box analytics to answer critical questions like: How is a checking account performing in the market? What’s the adoption rate of a promotional offer? How do customer transaction patterns vary across tiers?
• Cost-Efficient Transformations: With Databricks Autoloader, we seamlessly ingest billions of records from various file types (Avro, CSV) stored in Azure Blob Storage. This, combined with Serverless scaling, ensures that we can process data efficiently, without over-provisioning or incurring unnecessary costs.
• Ensuring Data Governance and Compliance: Unity Catalog enables robust data governance, allowing us to isolate single-tenant customer environments and maintain compliance with financial services regulations.

The impact for our banking customers is profound. The same tiering application that previously took a large bank two years to develop can now be deployed by their customers in just months. This is a significant time-to-market advantage that drives faster innovation and monetization.

“It’s exciting to see how Zafin is leveraging the Databricks platform to deliver data and AI-driven innovations that empower banks worldwide. By modernizing their infrastructure, Zafin is enabling financial institutions to unlock new opportunities for personalization, customer engagement, and growth-faster and more efficiently than ever before.”

— Junta Nakai, Vice President - Global Head of Financial Services, Cybersecurity and Public Sector GTM | Databricks

Looking ahead: Tapping into real-time innovations

In addition to enabling faster, more cost-effective data processing, the Databricks Platform has positioned us for future innovation. We’re exploring opportunities in AI and GenAI, such as real-time tiering insights, profitability modeling, and real-time offers with Databricks’ Delta Live Tables and streaming capabilities. To further streamline our analytics capabilities, we also plan to enable Lakehouse Federation for seamless integration across data sources.

As we continue to scale and innovate, we’re excited to unlock even more powerful capabilities for our customers—transforming data into actionable insights that drive the future of banking.

About Zafin

Founded in 2002, Zafin is a global provider of SaaS solutions for core banking modernization and transformation. Our award-winning platform enables banks to innovate their business models while modernizing their technology, ensuring transparency and fairness for banks and their customers. By enhancing operational efficiency, boosting revenue, and improving customer experiences, Zafin empowers financial institutions to achieve their strategic goals.

Zafin is headquartered in Vancouver, Canada, with a presence spanning offices and customers worldwide, including major banks like ING, CIBC, HSBC, Wells Fargo, Navy Federal Credit Union, PNC, and ANZ.

Learn more

Executives and Business Leaders can gain insights from peers at the Financial Services Forum–Shifting to Financial Intelligence, in New York City on Mar 20, 2025. Register today.

For technical professionals, visit the Data Intelligence for Financial Services webpage for more use cases and resources.