CASE STUDY:


From Legacy RPA to Agentic AI Transformation

Modernizing EMR Data Automation for a National Healthcare Network

Background

Healthcare networks must bring together data from multiple EMRs to manage population health and optimize operations under value-based care contracts. Accessing that data is not simple. APIs cover part of the workflow, but not everything. Historically, many organizations filled the gaps with scripted RPA bots that mimicked clicks and keystrokes inside EMRs.

While these bots delivered short-term productivity, they created long-term challenges. Scripts frequently broke after EMR updates, added technical debt, and introduced compliance and data integrity risks. According to Cloudmed/Becker’s research, only 7% of health systems describe their RPA deployments as mature despite years of investment. In practice, what started as a productivity tool had become a liability.

Problem:

The national healthcare network faced:

  • High administrative burden from brittle bots that required constant maintenance.
  • Data gaps where scripted RPA failed to capture information not exposed via APIs.
  • Compliance risk due to incomplete or non-transparent audit trails.
  • Escalating costs tied to ongoing reprogramming, IT support, and licensing fees.

Leaders sought a modern approach that could combine the resilience of APIs with the adaptability of AI, while eliminating technical fragility.

Solution:

The network partnered with Odesso to deploy a dual strategy:

Part 1: Middleware + APIs
Odesso implemented FHIR-based middleware that normalized EMR data and allowed structured updates, eliminating the need for bots to handle basic tasks like documentation status changes.

Part 2: Agentic AI Agents
To fill the gaps where APIs stop, Odesso deployed agentic RPA agents powered by LLMs and GenAI. These agents reason through unstructured notes, interpret clinical context, and act with human-in-the-loop oversight. Unlike scripted bots, agentic agents adapt when workflows or interfaces change, and they maintain transparent audit logs.

The combined approach allowed the network to:

  • Use APIs where possible for speed and security.
  • Deploy AI agents where data remained locked in unstructured formats or complex workflows.
  • Govern both layers through a single platform with dashboards, audit trails, and exception handling.

RESULT: Risks Converted to Rewards

  • Resilient data automation that survived EMR updates without reprogramming.
  • Complete population health data by blending structured APIs with AI-driven interpretation of unstructured inputs.
  • Audit-ready compliance with full logs of every agent action.
  • 70% reduction in support overhead compared to maintaining legacy RPA.
  • Scalable architecture now adopted across multiple states and EMRs.

Outcomes

Improved data completeness for VBC reporting and analytics.

Reduced IT burden, freeing staff from constant bot maintenance

Lower compliance risk with MEAT validation and transparent audit trails.

Extensible automation that can be applied to risk coding, quality abstraction, and prior authorization.

Financial ROI, as automation reduced manual effort and enabled fair reimbursement under risk-based contracts.

Odesso delivers healthcare automation designed for value-based care. Our Healthcare AI Platform combines APIs, middleware, and agentic AI to replace fragile scripted RPA with resilient, intelligent workflows. With capabilities spanning risk adjustment, quality measure abstraction, EMR data exchange, and population health analytics, Odesso helps provider networks modernize operations and succeed under value-based contracts.

CONTACT US