Protecting sensitive information in the healthcare industry
Orizon refined their data infrastructure and GenAI usage to serve their customers
Minutes to build code documentation
Of workflows automated with LLMs
BRL in productivity increases
Orizon, a health tech company within the Bradesco Group, enhances healthcare efficiency. Their platform connects insurance companies, hospitals, doctors and patients by using advanced analytics and artificial intelligence (AI) to detect financial fraud in medical billing. Orizon is dedicated to digital transformation, yet the company has struggled with fragmented data and inefficient manual processes — making it more difficult for their teams to combat fraud. Databricks provided a comprehensive solution for them, enabling data centralization, streamlining processing and automating manual tasks by 63%.
Facing inefficiencies in the healthcare documentation process
Orizon’s mission is to help prevent fraud in the healthcare industry by applying tools such as medical intelligence, analytics and automation. The forward-thinking health tech company had already begun leveraging their vast dataset encompassing hospitals, doctors and patients to fuel their innovation in AI. Yet the company was burdened by their internal data management, which hindered their fraud detection capabilities. Guilherme Guisse, Head of Data and Analytics at Orizon, explained, “We have around 40,000 medical rules in our system and add around 1,500 new rules each month. Every time a new rule was added, we would have a developer look at the code, document it and create a flowchart. It was a very manual and error-prone process.”
These rules — coded in legacy languages like C# and C++ — were necessary to determine a member’s healthcare eligibility but consumed considerable IT resources. Since adding a new rule took several days to complete, business analysts often had to request C++ developers to interpret code, which created even more bottlenecks for Orizon. Worse, it would require two developers in the business analysis team to build the documentation to enforce these rules.
To address these challenges, Orizon explored the potential of GenAI to automate code documentation and rule interpretation. Their first attempt enabled business users to query and receive instant explanations about rules by integrating a large language model (LLM) into Microsoft Teams. This innovation aimed to remove the developer bottleneck and significantly reduce the time required for updates. The envisioned LLM chatbot would provide accurate, real-time answers and graphical documentation, improving developer effectiveness and allowing teams to focus on company strategy and the product roadmap. By leveraging GenAI tools, Orizon aimed to transform their data management processes, enhance the accuracy and speed of their fraud detection efforts, and optimize the utilization of healthcare resources.
Modernizing data infrastructure for GenAI implementation
Orizon adopted the Databricks Data Intelligence Platform to more easily integrate all necessary components required for comprehensive data warehousing, processing and utilization. With a transition to a data lakehouse architecture, Orizon could consolidate their previously isolated databases into a single system, streamlining data workflows and ensuring efficient access and management. The inclusion of Delta Lake provided reliable data storage and management, enhancing the company’s ability to handle ACID transactions and unify streaming and batch data processing. This move laid the groundwork for scalable data practices, modernizing their legacy systems and unifying data, analytics and AI under one solution. This consolidation was pivotal in facilitating Orizon’s transition to leveraging GenAI in their internal (and eventually external) workflows.
The first task was embedding GenAI, along with its natural language processing capabilities, into Microsoft Teams so business users could get answers to questions without relying on developers. Using MLflow, an open source MLOps platform developed by Databricks, Orizon could seamlessly manage the entire machine learning lifecycle — including data ingestion, feature engineering, model building, tuning and experiment tracking — to ensure reproducibility. Once the Orizon team fine-tuned the Llama2-code and DBRX GenAI models with their own data and business rules, they securely deployed the models through Mosaic AI Model Serving, making them accessible and usable across the organization. Fortunately, the process was made even easier by Solutions Architects from Databricks. As Orizon continued to incorporate GenAI into other internal workflows, their initial use of LLMs for text and graphical documentation of source code saved time and effort, benefiting business users on the product and commercial teams.
Given Orizon’s presence in the healthcare industry, their data strategy required enhanced data governance and security. Unity Catalog ensured secure data management of business rules, which were proprietary to the company and vital to training the new LLM. Guisse expanded, “We couldn’t send this data outside the company because our business rules are confidential. To develop our LLM securely within Orizon, we needed a partner like Databricks to empower our users to create custom models based on our own data.” Orizon required strict governance to protect sensitive information inherent in the healthcare industry in a location where fraud was particularly rampant. Thanks to Unity Catalog’s ability to define and enforce granular permissions, Orizon could maintain the necessary security and compliance, ensuring that sensitive data remained protected yet accessible to authorized personnel for critical operations. This ultimately improved their ability to fight healthcare fraud effectively.
Maximizing cost savings with optimized resource utilization
To say Orizon has experienced greater operational efficiency since implementing Databricks is an understatement. According to Guisse, “With our brand-new GenAI features, we are currently processing 63% of tasks automatically. So, our developers can focus on development, not just documentation.” Since the health tech company is already experiencing substantial improvements, it has freed up one and a half developers, who can focus on high-value tasks such as implementing new fraud detection rules. This shift has also translated to significant cost savings for Orizon — approximately $30K per month in better-used resources. Even more impressive, the documentation process takes less than five minutes, dramatically accelerating workflows, with the potential to increase the rate of new rules from 1,500 to 40,000 per month — translating to 1 billion Brazilian reals (BRL) in added productivity.
Furthermore, Databricks’ advanced data strategy has transformed data governance and security at Orizon, crucially aiding their mission to combat healthcare fraud in Brazil’s tough market. Orizon is proud to adhere to stringent privacy regulations, safeguarding sensitive patient and medical billing information. This robust approach empowers Orizon to leverage historical data and AI models to evaluate and verify medical procedures, significantly reducing fraud, contributing to cost savings for insurance plans and improving healthcare access and quality for the people of Brazil.
Orizon can now explore additional innovative AI use cases. Next, the team hopes to develop a large language model to check and validate the materials approved for use in medical and hospital procedures. It will ensure that only the necessary and appropriate materials are used, reducing costs and improving the quality of care. After this initiative, Orizon wants to leverage AI to enhance customer service by reading application logs and answering customer inquiries in real time. Currently, it takes the customer service team five to six days to respond to inquiries. However, with the help of GenAI, they’re hoping to expedite the process to foster customer satisfaction and loyalty. These continued advancements are part of Orizon’s commitment to using cutting-edge AI technologies to improve healthcare delivery in Brazil.
