• 

  AI-governed healthcare operations dashboard

ACEmed — AI-Governed Healthcare Operations

AI-Governed Healthcare Operations, Built for Human Care — Reducing Referral Delays with Healthcare AI You Can Audit and Trust.

Inspiration

Healthcare systems often struggle with referral delays, incomplete documentation, scheduling inefficiencies, and fragmented coordination between systems and teams. While many healthcare AI solutions focus on diagnosis or treatment, we saw an opportunity to improve the operational layer of healthcare instead.

ACEmed was inspired by a simple question:

What if healthcare organizations had an AI operations teammate that could coordinate referrals, check completeness, support administrative workflows, and remain fully human-governed?

Rather than replacing clinicians, ACEmed is designed to support healthcare teams by reducing friction, surfacing missing operational information, and helping organizations move patients through care pathways more efficiently.

What it does

ACEmed is an A2A-compatible healthcare operations agent built for the Prompt Opinion ecosystem.

The system reviews synthetic referral context, checks operational completeness, applies governance and workflow logic, and provides audit-safe recommendations for the next operational step.

Examples include:

• Referral intake review
  
• Prior authorization completeness checks
  
• Scheduling and administrative coordination
  
• Human-in-the-loop escalation recommendations
  
• Audit-safe summaries and operational handoffs
  

Importantly:

ACEmed does not diagnose or recommend treatment.

Instead, it focuses on healthcare operations and coordination, helping reduce delays and unnecessary administrative back-and-forth.

How we built it

We used the Prompt Opinion ecosystem and external A2A agent architecture to create an interoperable healthcare operations assistant.

The solution combines:

• A2A interoperability for agent communication
  
• Generative AI for operational reasoning and summarization
  
• Synthetic healthcare data for safe testing
  
• Human-in-the-loop governance for safety and trust
  
• Audit-safe operational outputs for accountability
  

The project was designed to align with healthcare interoperability principles and can evolve toward FHIR-context-aware workflows using SHARP-compatible context propagation.

Challenges we faced

One of the biggest challenges was balancing:

AI capability vs healthcare safety

We deliberately constrained ACEmed to operational support rather than clinical decision-making.

Another challenge was ensuring interoperability while remaining practical within hackathon timelines. We focused on building a solution that could realistically operate in healthcare settings today while remaining lightweight enough to demonstrate quickly.

What we learned

We learned that healthcare AI does not always need to focus on diagnosis to create meaningful impact.

Operational friction is a major pain point in healthcare systems, and thoughtfully governed AI can help reduce delays, improve coordination, and support care teams without replacing human oversight.

We also learned the importance of interoperability standards such as A2A, MCP, SHARP, and FHIR for building future-ready healthcare systems.

What's next for ACEmed

The long-term vision for ACEmed is to evolve into a governed healthcare operations layer capable of coordinating multiple healthcare workflows and specialist agents.

Future directions include:

• FHIR-native integration
  
• SHARP context propagation
  
• Multi-agent healthcare coordination
  
• Expanded operational workflows
  
• Advanced audit and governance tooling
  

The future of healthcare AI is not isolated systems.

It is interoperable, governed, human-centered collaboration.

Built With

• a2a
• fhir-ready-architecture
• gemini-api
• generative-ai
• google-adk
• governance
• healthcare-ai
• human-in-the-loop
• ngrok
• node.js
• prompt-opinion
• react
• sharp-extensions
• synthetic-data
• tailwind-css
• typescript
• vite