How Inova modernized its data architecture to prepare for AI apps
Inova Health System Accelerates AI Integration Through Data Modernization
Northern Virginia-based Inova Health System has transitioned over 70 artificial intelligence applications from experimental phases to live production by compressing a four-year data modernization roadmap into six months. The nonprofit health system, which records more than 4 million patient visits annually, partnered with data integration firm Fivetran to standardize its clinical, imaging, and financial data models, according to a recent Fivetran case study.
How Inova Operationalizes AI Across Clinical Workflows
Inova does not aim to replace third-party developers but rather to integrate specialized AI capabilities into its existing digital portal. According to Jon McManus, Inova’s chief data and AI officer, the goal is to provide a “wayfinding” experience for clinicians, preventing them from having to manage 100 different, disconnected AI interfaces.
Current clinical deployments include:
- Ambient Scribing: Over 1,000 providers use Abridge to automate documentation during outpatient and emergency visits.
- Diagnostic Imaging: Algorithms identify microclotting in chest CT scans for post-stroke patients.
- Administrative Efficiency: A pilot project with Elevance Health uses AI to pre-chart patient visits, which may expedite insurance approvals by ensuring necessary clinical information is captured in real-time.
Why Guardrails and Vendor Diversity Matter
To avoid “vendor lock-in,” Inova deliberately maintains a diverse technology stack rather than relying on a single model developer. McManus notes that this strategy allows the health system to protect its proprietary business logic—the “brain” of its software—while remaining flexible enough to swap out components as AI technology evolves.
Safety remains the primary metric for all deployments. Every AI agent and model is measured against whether it increases patient safety, equity, or access. When providers use AI scribes, they are required to inform patients that the encounter is being recorded, a policy intended to maintain transparency and trust, according to McManus.
Future Trends in Healthcare AI
The shift toward “agentic AI”—systems that can perform tasks like referring patients to out-of-network specialists—represents the next frontier for health systems. By sending data back to internal clinicians, these agents allow Inova to maintain oversight even when care is delivered by community partners. This proactive, background-oriented approach, such as AI tools that identify open appointment slots for access teams, suggests a move toward AI that manages the “business of medicine” without requiring constant human input.
Frequently Asked Questions
How does Inova prevent AI bias and ensure patient privacy?
Inova enforces a strict validation bar for every vendor, judging tools against safety and equity standards. Patient privacy is protected by clear communication during exams and by utilizing governed data pipelines that prevent data mishandling.
What is the benefit of a “data modernization” effort for AI?
Modernization allows health systems to break down silos between clinical, laboratory, and financial platforms. By using tools like Fivetran and DBT, Inova created a single, governed ingestion model that allows AI apps to access consistent information across the entire organization.
Is Inova building its own AI from scratch?
No. McManus explains that Inova incorporates third-party capabilities from providers like OpenAI and Anthropic, often building smaller, custom features on top of those models to maintain control over the user interface and specific clinical workflows.
Are you navigating a digital transformation in your own healthcare organization? Share your experiences in the comments below or subscribe to our newsletter for more insights on the future of health IT.