Investing in smarter personalization with advanced analytics
TTB migrates to the Databricks Platform to enhance customer experience and security
Increase in highly targeted offer
Faster digital sales growth
Customers reached with personalized products and services
TMBThanachart Bank (TTB), one of Thailand’s top five banks, is at the forefront of financial innovation, serving millions of customers with personalized banking solutions. However, their legacy data infrastructure constrained their ability to scale personalization and meet growing customer demands. Challenges like high data latency, limited scalability and system inflexibility made delivering timely insights and combating fraud increasingly difficult. Partnering with the Databricks Data Intelligence Platform, TTB migrated their in-house personalization engine Delphi to a unified, scalable data architecture, unlocking real-time analytics and AI-driven capabilities. Today, TTB reaches 5.5 million customers with hyperpersonalized services, delivers targeted offers at 10x the previous capacity and has doubled digital sales growth — all while enhancing fraud prevention and operational efficiency. The transformation not only resolves past challenges but also sets the stage for continuous innovation in customer engagement and financial security.
Data latency hinders personalization and fraud prevention in banking
TMBThanachart Bank (TTB) is a leader in the financial sector with a strong “digital-first strategy.” Dedicated to providing innovative financial solutions, TTB serves both retail and commercial markets through a comprehensive portfolio of services, including personal loans, mortgages, credit cards and hire purchase offerings. Their mission is to empower the financial well-being of Thai individuals and businesses by leveraging data and technology to deliver hyperpersonalized experiences, enhanced security and seamless banking solutions.
“At TTB, our goal is to bridge the gap between customer expectations and financial realities through tailored solutions powered by data and AI,” Naris Sathapholdeja, Chief of Data and Analytics Group at TTB, said. “Our commitment to innovation ensures that we are not just meeting customer needs but also anticipating them.”
TTB’s forward-thinking approach focuses on creating a “Segment of One” strategy, powered by their built-in-house Delphi Personalization Engine. Delphi leverages advanced analytics and AI to tailor financial services to each customer’s unique needs. Key use cases include personalization to deliver individualized recommendations for credit cards, loans and spending adjustments; fraud detection and security to combat mule accounts and digital fraud through real-time analytics and detection; propensity scoring to identify the right products for the right customers at the right time; risk management and compliance to ensure robust anti–money laundering measures and regulatory adherence; and customer engagement to enhance satisfaction and loyalty through omnichannel marketing and AI-driven insights.
Unfortunately, their on-premises data infrastructure limited scalability, restricting personalized products and services to just 500,000 customers. Batch processing caused high data latency and delays in insights, due to a two-day lag time and inability to handle streaming data. Shared resource constraints heightened the risk of system breakdowns, increasing the reliability risks of their data operations. Numerous dependencies and limited conditions made their data architecture extremely inflexible, slowing the deployment of new features.
“We reached a tipping point where our legacy architecture couldn’t keep up with the growing complexity and volume of data,” Wipada Chanthaweethip, Head of DataOps at TTB, explained. “It became clear that, to lead in innovation and customer satisfaction, we needed a scalable, cloud-based solution.”
These challenges led TTB to partner with Databricks to migrate their data, including Delphi, to a unified lakehouse architecture. This decision allowed TTB to address their core limitations while laying the foundation for scalable and efficient data analytics.
Empowering exceptional customer experiences with the Databricks Data Intelligence Platform
Migrating Delphi to the cloud via Azure Databricks provided TTB with a more adaptable and scalable environment. The integration with existing cloud services, such as Google Cloud Pub/Sub and Azure Event Hub, enabled a more streamlined data flow and reduced dependencies on legacy systems — unlocking new capabilities to support their hyperpersonalization, fraud prevention and advanced analytics initiatives. “Databricks provided the scalability and flexibility we needed to bring our Segment of One strategy to life,” Naris noted. “With a unified data architecture, we can seamlessly integrate analytics and AI into every aspect of our operations, empowering us to deliver meaningful customer experiences at scale.”
Centralizing their data operations on a single lakehouse architecture allows for real-time processing and supports both structured and unstructured data, ensuring a cohesive environment for analytics and AI-driven decision-making. “It bridges the gap between legacy batch processes and the real-time capabilities required for advanced personalization and fraud detection,” Wipada explained.
Governance was a top priority for TTB as they expanded their use of analytics. By adopting Unity Catalog, TTB established centralized data governance across teams and partner ecosystems, ensuring secure and efficient data access. “Unity Catalog enables us to collaborate across departments while maintaining strict compliance and security protocols,” Wipada added. “It’s the backbone of our ability to scale AI and analytics safely.”
To orchestrate and automate complex data workflows, TTB leveraged Databricks Workflows, enabling efficient, end-to-end data pipeline management. Workflows streamlined the integration of data from diverse sources, enhanced operational reliability and supported the rapid deployment of AI models into production. “With Databricks Workflows, we’ve significantly reduced the time and effort required to manage data pipelines, allowing our teams to focus on delivering innovative customer solutions,” Wipada said.
With the Databricks Platform, TTB enhanced their ability to process real-time data, enabling Delphi to deliver hyperpersonalized recommendations and identify fraudulent activities as they occur. The platform’s support for machine learning models allowed TTB to refine their propensity scoring and risk management use cases with greater precision. Naris explained, “Real-time capabilities are critical for fraud prevention. With Databricks, we can act immediately, protecting our customers while minimizing disruption.”
Doubling digital sales with meaningful, targeted offers
By adopting the Databricks Platform, TTB transitioned from a constrained infrastructure to a scalable, real-time and reliable platform. The results have been transformative.
Over the course of a yearlong engagement, the Databricks team worked closely with TTB’s management, champions and practitioners to align on a new data architecture that could support bankwide analytics. This collaboration resulted in the migration of over 1,000 tables from their legacy on-premises system to Azure Databricks. This migration drove a 38x year-over-year increase in data consumption. Additionally, TTB achieved 100% adoption of Unity Catalog, ensuring consistent, robust data governance across the bank and their partner ecosystem.
Personalized products and services now reach 5.5 million customers, a dramatic increase from the previous 500K limit. Insights are near real time, reducing delays from two days to just one day (or less) for most sources, while enabling streaming data processing. Dedicated environments have minimized system breakdown risks, ensuring stability and operational excellence. Meanwhile, dependencies and constraints have been replaced with a flexible infrastructure, enabling faster deployment of new features — significantly reducing time to market.
With Databricks powering the Delphi Personalization Engine, TTB achieved a 10x increase in card generation capacity, enabling the delivery of highly targeted offers to millions of customers. This contributed to a dramatic 2x acceleration in digital sales growth, including increased adoption of credit cards, deposits and other financial products.
The Delphi Personalization Engine now processes 4.3 billion data points daily and 20 million real-time responses, enabling hyperpersonalized experiences that resonate with customers. As Wipada explained, “Our personalization efforts have transformed how we engage with customers. With more than 18 million personalized messages delivered daily, we’re seeing increased satisfaction and loyalty.”
“Our journey with Databricks has only just begun,” Naris said. “We’re building a future where data and AI power every aspect of our operations, ensuring we stay ahead of customer expectations and industry trends.”