Bupa Australia is part of the Bupa Group, a multi-billion dollar international healthcare company serving over 60 million customers worldwide. Globally, Bupa Group has 47.1 million health insurance customers, 20.7 million health provision customers, 520 health clinics, 52 Mindplace health centres, 28 hospitals and over 900 dental centres.
Bupa Australia helps customers live longer, healthier, happier lives through healthcare services spanning health insurance, dental, optical, medical centers, mental health clinics, aged care, and digital health. As the company pursued personalized care across these services, fragmented legacy platforms made it difficult to govern sensitive health data or deliver a single source of truth. By standardizing on the Databricks Platform with Unity Catalog governance, Bupa consolidated more than 90% of business data onto a single governed platform, accelerated analytics delivery, and laid the foundation for AI-assisted, personalized health plans.
Personalized care needs a trusted data foundation
Bupa Australia's ambition to deliver personalized, preventative care depends on having a unified view of each customer across every service they touch, from a dental checkup to a hospital claim to a telehealth consultation. That single view enables the predictive models and digital health twins now central to the company's strategy. But that unified view had to be built from scratch. As Bupa grew through acquisitions over decades, each business unit had developed its own technology independently. “We had multiple data systems and legacy platforms, each with its own complex governance," said Ed Falconer, Chief Data Officer of Bupa Australia. "The same metric could exist in different tables and return a different answer."
The fragmentation showed up in everyday work. In team meetings, something as straightforward as a customer count would spark a debate because finance had pulled the number from one system and the data team from another. “The question was always the same,” Ed said. “Where is the single source of truth?”
As Bupa leaned into customer-centric personalization, those gaps put more strain on the underlying infrastructure. The personalization team needed roughly 1,500 data points to power decisioning across millions of customer touchpoints each year, but legacy systems slowed every step. Some data loads arrived only once a week, limiting how quickly teams could act on emerging trends. Complex compute processes could take more than 24 hours to run across millions of records. The ambition was there. The infrastructure was not.
How did Bupa build a centralized, governed data platform on Databricks?
Bupa’s answer was a platform decision, not a point solution. The organization chose to consolidate all data, analytics, and machine learning onto the Databricks Platform, with a mandate to decommission legacy systems over time. “There is one system that we’re going to build this analytical data environment on, and it’s going to be on Databricks,” Ed said.
Over a three-year program, the data team built the Bupa Data Platform (BDP) and a dedicated operations group was formed to keep it running around the clock while migration and decommissioning continued in parallel. Execution mattered as much as architecture. The team moved away from waterfall delivery to a rapid test-and-learn model in which engineers continuously built, validated, and shipped.
Customer privacy and data governance come first
Because BDP brings together sensitive customer data, governance had to be foundational rather than an afterthought. Unity Catalog provides centralized access control, and Bupa layers additional protections on top. The governance team works alongside engineers to ensure that data access, usage, and sharing meet Bupa's privacy standards at every stage. “Maintaining privacy and governance come first. Every decision downstream depends on understanding exactly what data you're handling and the permissions our customers have granted,” said Ed.
By integrating transformation and governance into a single platform, Bupa built the trusted foundation on which its use cases depend. "Customer consent is crucial in all discussions," Ed said. "The treatment and concern about using data without the permission of the customer is absolutely paramount. Everything else plays out from there." Each of the capabilities below draws on governed data within BDP, operating within the boundaries of customer consent and the privacy safeguards that Unity Catalog makes possible.
Governed customer 360
To power personalization, Bupa first needed a single, trustworthy representation of each customer. Within BDP, the data team integrated data from across the business into a single, governed view per customer, representing a digital health twin. That unified profile gives clinicians, analysts, and decisioning tools the same answer to the same question, replacing the fragmented records that once lived across disconnected systems.
Personalization engine
Bupa's personalization team sources roughly 1,500 data features directly from BDP and connects them to campaign management and decisioning tools, powering millions of customer touchpoints per year. Before the migration, the team was encumbered by legacy systems that limited the number of campaigns it could run and the speed of iteration. With features now served from a single governed platform, the team can launch more campaigns, experiment at a higher rate, and respond to behavioral and health signals while they are still actionable.
Predictive health analytics
With a complete customer history in one place and with the consent of its customers, Bupa layers machine learning models on top to identify patterns across a customer's past and current state. These models can surface emerging risks, such as indicators of prediabetes, and inform personalized health plans that anticipate future needs rather than simply react to past events. The predictive capability is central to Bupa's shift toward preventative, personalized care.
Bupa's results on Databricks: 8x faster analytics at scale
After consolidating onto the Databricks Platform with Unity Catalog governance, Bupa now holds more than 90% of all business data in a single governed environment and delivers data and analytics outcomes over 8x faster than 3 years ago.
Today, more than 90% of all business data resides in BDP, creating a shared foundation the business can trust. The most immediate impact has been speed. Bupa now delivers data and analytics outcomes over eight times faster than it did three years ago. Many data loads have moved from weekly to daily, enabling teams to respond to health and behavioral signals while they are still actionable.
The personalization team has seen the most tangible operational change. The group sources its roughly 1,500 features directly from BDP and connects them to campaign management and decisioning tools, enabling more campaigns, faster iteration, and greater experimentation across millions of annual touchpoints. Bupa has identified double-digit millions of dollars in opportunity from the transformation and expects further incremental growth as personalized offerings scale. “Our personalization team went from being encumbered by legacy systems to running thousands of campaigns a year directly from the platform,” Ed said. “That speed changes what’s possible.”
One platform replaces decades of fragmentation
Bupa’s legacy environment consisted of multiple disconnected data warehouses, analytics platforms, and reporting tools accumulated through years of acquisitions. Maintaining these systems meant complex governance, inconsistent metrics, and weekly data loads that delayed decision-making. The Databricks Platform replaced this architecture with a single governed environment, eliminating cross-system ETL, consolidating access controls under Unity Catalog, and providing a modern analytical toolkit that the legacy systems could not match.
The governed foundation is now the launchpad for what comes next. Bupa’s data-driven, personalized and preventative health plans are in active development this year, powered by increasingly sophisticated predictive models that leverage a growing range of signals and predictors to anticipate future health risks. As with all data flowing into BDP, signals are incorporated only with explicit customer consent, a principle Bupa treats as non-negotiable regardless of the data source.
The team is also experimenting with Genie and exploring how agentic AI can operate within governed workflows, with humans directing systems of decisions managed by AI. "Customer centricity starts with truly knowing your customer, and that requires a single, trusted source of truth," Ed said. "Databricks gave us that foundation. Everything we build from here starts there."