Orizon helps protect millions of Brazilians from unnecessary risk and waste in healthcare. With over 220,000 providers and 13,000 pharmacies, managing complex and unstandardized claims data in PDFs, Word documents, and spreadsheets was error-prone and costly. Facing long deployment times and fragmented infrastructure, Orizon turned to Databricks Data Intelligence Platform to automate claims, improve accuracy and deploy AI solutions faster to reduce waste and ensure correct reimbursements for patients.
Adapting healthcare operations to reduce risk and waste
Few know the challenges of healthcare like Orizon. As a healthcare technology partner to hospitals, clinics and pharmacies, Orizon manages some of the most sensitive and critical patient data. Delivering accurate care and outcomes depends on precise coordination of everything from medical claims and regulatory approvals to patient-specific procedures—down to the exact screw needed for a knee surgery—across 26 million patients and 220,000 providers.
When data is inaccurate or delayed, the consequences ripple across the system. Lost revenue from incorrect or denied claims, higher financial burdens on patients and provider frustration that strains relationships are just some of the ways bad data can impact the business and patients. For Orizon, efficiency isn’t just operational—it’s the difference between protecting margins, preserving trust and enabling better patient care.
Intending to minimize waste and optimize care, Orizon is always looking for ways to maximize accuracy and efficiency. However, while the healthcare industry constantly works to optimize care, administrative details like forms and document formats can take countless forms and use various technologies. So while data volume alone could be a challenge, issues around data consistency added even more complexity for Orizon. “There was no standardization across documents, with many images and different formats like PDF, Excel, X-rays and Word,” described Rodrigo Rodriguez, Advanced Analytics Coordinator at Orizon. “Just reading the documentation was hard enough. But we also needed to analyze the data and extract the right information.”
Increasingly, the team saw that the path forward would require the ability to ingest these data streams in real-time and turn them into accurate, actionable information. And for that, the team turned to the Databricks Data Intelligence Platform.
Unlocking agentic AI to automate claims operations
Databricks offered a unified platform to manage the entire data and AI lifecycle — from ingestion to model deployment — enabling Orizon to scale AI confidently. “Grounding our data in AI is critical, as it’s dealing with healthcare claims that directly impact patients,” said Rodrigo.
Using Delta Lake, Orizon consolidated structured and unstructured data into a single source of truth to reduce disputes between providers and payers, while Delta Sharing allowed seamless collaboration across teams and providers. Lakeflow Jobs automated the ingestion and processing of massive datasets, transforming raw data into actionable insights to reduce waste. Unity Catalog ensured fine-grained governance over sensitive patient information, safeguarding patient trust while meeting Brazil’s strict LGPD compliance.
Where Orizon saw the most significant impact was Mosaic AI agentic solutions, which accelerated the development of agentic AI solutions—directly reducing operating costs and enabling new patient-facing services. With Databricks, Orizon could pair lightweight models for routine claims classification with more advanced GenAI models for regulatory chatbots like Nori and Pietra, matching the right tool to the job. That flexibility kept costs down on high-volume tasks while speeding delivery of new AI assistants. Tasks that previously took months, such as training and deploying a model to analyze claims or translate technical jargon for patients, can now be completed in weeks. “Before working with Databricks, deploying AI models into production was very hard. One model took nearly three years to deploy. Now we can ship a GenAI chatbot to solve a customer challenge in weeks,” explained Rodrigo.
This agentic AI solution classifies whether a medical claim needs to be referred for human audit or can be automatically approved — enabling faster, more accurate decision-making for providers and patients alike. Orizon has also introduced a growing set of AI specialists powered by Mosaic AI. Nori, an AI regulatory compliance expert, gives employees immediate answers on complex compliance regulations; Pietra, an AI information security advisor, helps teams adhere to LGPD (Lei Geral de Proteção de Dados — Brazil’s General Data Protection Law) and privacy best practices while flagging risks; and Gisele, an AI supply management specialist that extracts and interprets details from complex medical documents to improve sales efficiency and reduce costs. Together, these solutions are helping Orizon boost productivity, cut costs, and deliver safer, faster outcomes across its healthcare ecosystem.
Driving measurable impact and better patient care
Databricks has transformed the way Orizon operates, enabling faster, more cost-efficient workflows. Automating claims audits increased productivity by 40%, saving R$1.5 million (over $280,000 USD). Supply management has also been streamlined with AI, automatically extracting critical information from highly complex medical documents. This solution improved sales efficiency and reduced costs by an additional R$1 million, bringing total savings to R$2.5 million (over $460,000 USD) while ensuring standardized forms are 100% complete for accurate reimbursements and faster approvals.
Crucially, Orizon achieved these results without compromising on governance. By bringing the model to the data under Unity Catalog, they avoided data egress, stayed compliant with LGPD, and protected patient trust while still innovating quickly.
AI-powered chatbots now provide concise, real-time explanations for denials, regulatory questions, and provider requests. What once required hours or days of manual review is resolved in just three hours, enabling the team to handle five times more weekly requests. Natural language outputs further simplify complex rules into actionable guidance, helping patients and providers resolve issues instantly while reducing administrative delays.
“Databricks has been critical to our business, helping us reduce costs, simplify operations and deploy models faster than ever,” said Rodrigo. “The platform also lays the foundation for future agentic AI solutions that will continue fighting waste and scaling impact across millions of patients.”