How FinThrive Uses Databricks AI/BI Genie to Cut Query Time from Days to Minutes

FinThrive holds a dominant position in US healthcare revenue cycle management, serving more than 1,200 hospitals and health systems including some of the largest in the country. Its life sciences division handles a different kind of demand: pharmaceutical companies and channel partners who need access to de-identified real-world healthcare data to inform clinical development, market strategy, and population health research. That data sits in a petabyte-scale lakehouse built on Databricks, containing over 100 terabytes of de-identified real-world data organized in Gold-layer materialized views.

The bottleneck was the SQL dependency. Every time a pharmaceutical company submitted a feasibility request—asking about specific patient populations, therapeutic areas, or longitudinal data quality—a small team of analysts had to write custom SQL queries from scratch. Each request had different logic and different criteria. Different analysts would approach similar queries differently, producing inconsistent answers and creating variability that eroded customer trust. The typical turnaround was 3 to 5 days. In a competitive market where channel partners shop multiple providers, that delay was losing deals.

FinThrive deployed Databricks AI/BI Genie on top of their existing lakehouse architecture and Unity Catalog governance layer. The implementation leverages materialized views built on Gold-layer tables, enabling sub-second query response even against datasets with billions of records. The team manages data security by providing only aggregated, HIPAA-compliant access through Genie, ensuring no detailed patient information is exposed. The natural language interface then sits on top of this architecture, letting users phrase queries in plain English and receive structured results in minutes.

The impact was immediate for the sales team. Representatives now input customer criteria directly into AI/BI Genie and get responses back to customers within hours rather than waiting days for analyst queue availability. Consistency improved dramatically—the same question asked twice now returns the same answer, regardless of who asks it. Analysts, freed from repetitive feasibility queries, shifted to higher-value analytical work. FinThrive's data shows query response time fell below 24 hours, down from the previous 3-to-5-day standard.

The next phase extends Genie access to marketing teams and, eventually, external channel partners through a secure application layer built on the AI/BI Genie API. Integrations through Microsoft Teams are in development for internal users. FinThrive's architecture positions the lakehouse as the single source of truth for both operational and analytical queries, and Genie as the interface through which the entire organization—and eventually external partners—will interact with that data without needing SQL expertise.