Product descriptions:
IndusInd Bank serves 41 million customers across India through a vast network of 3,000 branches and digital channels, driving the country’s shift toward modern, technology-enabled banking. As their operations expanded, the bank’s legacy data warehouse was unable to keep pace. Fragmented data, limited observability and slow refresh cycles made it difficult to deliver timely insights or scale AI initiatives. By migrating to the Databricks Data Intelligence Platform, IndusInd established a unified enterprise data platform (EDP) that integrates 1.5 petabytes of data across 44 business areas, enabling real-time reporting and advanced analytics for thousands of users. What once took days now happens in minutes, and what was once fragmented is now fueling innovation — from hyper-personalized customer experiences to AI-driven decision-making that redefines how the bank operates and serves their customers.
Clearing the path to AI-driven banking
IndusInd Bank, the fifth-largest private sector lender in India, has been committed to delivering digital transformation, innovation and customer-centric services to the people of India. To deliver frictionless, personalized experiences to more than 4.1 million customers, the bank invested in data and analytics as the foundation for their next wave of growth. “We want to be the AI-first bank. And given that requirement, our data lakehouse plays a major role in supporting real-time engagement and hyper-personalization for our customers,” Shiv Kumar Bhasin, Chief Transformation Officer at IndusInd Bank, said.
To enable near real-time analytics and faster, data-driven decisions, IndusInd needed a unified view of customer data and stronger enterprise-wide reporting. The solution: migrate to Databricks to build a single EDP in the cloud, featuring real-time data streaming from source systems, built-in data quality checks at the time of data ingestion and data governance frameworks integrated to drive generative AI-driven machine learning (ML) models for data science teams across the enterprise.
Today, the EDP supports more than 1,200 reports and dashboards for finance, HR and operations. Automated reporting powers account management and HR insights, while behavioral analytics enhance customer engagement and drive informed decision-making. Data from contact centers and third-party providers inform models for personalized nudges and social media insights feed campaign automation.
The legacy Azure Synapse warehouse was unable to scale for AI/ML workloads or handle unstructured data. “We did not have a comprehensive data approach. Whenever business users were carrying out analytics, they used to miss out on certain parts of the data,” Shiv recalled. Without a unified data model or observability tools, data quality issues went undetected, leading to a reliance on slow, daily or multi-day refreshes. Integration across 95 operational systems and 4,000 datasets spanning 30 departments was unmanageable. “With our old warehouse, data ingestion took one or two days. We needed to bring the data into our lakehouse in near real time,” Shiv explained.
The bank realized that achieving their AI-first vision required a modern, unified platform capable of handling both historical and streaming data while enforcing governance, trust and scalability.
Building an AI-ready enterprise data platform on Databricks
The migration to Databricks unified IndusInd’s data and created a single, governed source of truth. “We are the only bank in India that uses a single data lakehouse on a single technology, which is Databricks,” Shiv said. The platform encompasses 44 business areas, comprising over 500 tables and 33,000 data elements, ensuring modularity and ease of maintenance.
The architecture decouples compute and storage for scalable, cost-efficient operations. Databricks ingests data from over 120 structured and unstructured sources, including contact center audio and video, integrated into a unified model for operational reporting and ML workloads. “Unlike our previous data warehouse, Databricks lets us bring in a lot of unstructured data. Our contact center transformation is going to get fueled by this data lakehouse,” Shiv noted. Alation integration enhances discoverability, enabling teams to trust and explore data more quickly.
At the core of the EDP is Delta Lake, which provides a medallion architecture to refine and organize data through three stages. Bronze tables capture raw data, Silver tables transform and standardize it and Gold tables deliver curated data for analytics and AI. This layered design has made the bank’s analytical models more robust and scalable. MLflow helps manage the lifecycle of the EDP’s ML models. It supports versioning, experimentation and deployment across the digital channels where real-time analytics are critical.
To ensure security and compliance, Unity Catalog enforces role-based access controls across every layer of the data stack. Data trust is further reinforced through 2,800 active data quality rules and continuous data observability dashboards that monitor accuracy and timeliness in near real time. Databricks has advanced data privacy controls, including automated identification, masking and encryption of over 2,000 personally identifiable information (PII) elements to meet internal security standards.
Databricks also improves cost control and performance. Real-time workload monitoring enables adjustments in seconds. “Databricks has an excellent FinOps framework built in. Minute by minute, we can observe what is being consumed and where to focus our engineering efforts,” Shiv explained. Serverless compute has boosted performance and cost efficiency, with further migrations planned.
Built entirely on Microsoft Azure, the EDP integrates with native Azure services while Databricks handles processing, governance and analytics. Power BI dashboards provide real-time insights to business users, while Solace middleware and Databricks streaming connectors enable low-latency ingestion for operational decision-making.
By unifying their architecture on Databricks, IndusInd established a modern, cloud-native foundation that powers real-time decisioning, regulatory compliance and AI-driven customer experiences — all on trusted, high-quality data.
Delivering real-time financial decisions for customers
The shift to Databricks transformed how IndusInd Bank manages, accesses and acts on data. The EDP now supports over 1,200 operational reports and dashboards, empowering teams across 44 business areas with timely and trusted information. Top dashboards, tracking financial indicators such as branch performance, are accessed daily at more than 3,000 locations. Real-time insights now drive critical operations across the enterprise.
Before Databricks, data availability lagged by up to two days. Today, near real-time analytics enable faster, smarter decision-making. “Going from one or two days to 45 minutes or less is a significant leapfrogging,” Shiv said. This acceleration supports compliance, such as RBI MSME loan application requirements, which were implemented in just 12 weeks.
The 15-month migration, from engineering design to user onboarding, exceeded expectations. “Most of the leading SIs [system integrators] told me it was a two-and-a-half-year project. I challenged my team to do it faster, and they did it in 15 months, which is a major success,” Shiv added. Approximately 75% of users have transitioned to Databricks, with full adoption expected within the year.
Looking ahead, IndusInd is exploring generative and agentic AI to enhance customer experiences, including voice-enabled interactions, contact center modernization and branch knowledge assistants powered by real-time behavioral analytics. “Databricks is the single best platform in the market. It handles the complexity of data and provides the flexibility for future use cases that require GenAI-driven architectures,” Shiv said. With the Databricks Platform as the foundation, IndusInd Bank continues to expand their use of tools like Delta Sharing and Lakeflow Spark™ Declarative Pipelines (formerly Delta Live Tables) to enable broader collaboration and automation. The cloud-native EDP positions the bank to deliver the next generation of intelligent, customer-centric banking.