Case studies

Medical

Medical – Regional Urgent Care Network: Intake-to-Disposition Orchestration

Timeline: Build: 9 weeks. Deploy: 3 weeks per wave. Optimization: 10 weeks protocol alignment.

Business overview

Industry: urgent care. Size: 11 sites. Complexity: variable staffing, seasonal surge. Systems: Epic, CRM for employer contracts.

Common pain points

Fragmented intake, triage overload, inconsistent discharge. Manual bottlenecks at peak flu season. Scale stress broke phone queues.

Full AI automation system (PrimeAxiom solution)

Intake Agent structures complaints and history; Triage Assist suggests ESI-aligned pathways (non-diagnostic); orders staged for provider approval; Discharge Agent builds education and PCP follow-up.

Automation workflow (detailed system flow)

1. Trigger: online/kiosk/on-site.
2. AI maps symptoms to structured fields.
3. FHIR resources staged.
4. Provider edits; protocol rules fire.
5. Orders, scripts, work notes; SMS instructions.
6. Provider attestation.
7. Quality metrics refine prompts.

AI agents & logic

Intake NLP Agent, Acuity Assist Agent, Discharge Content Agent, Follow-Up Scheduler Agent.

Technical architecture

BAA-covered hosting, Epic integration, auditable LLM prompts, PostgreSQL event store.

Tools & integrations

Epic MyChart APIs, Twilio, PostgreSQL, OpenAI, location services

Results

Door-to-provider time −23%; duplicate data entry −68%; post-visit ED bounce-backs −15% in pilot sites; nursing admin time −120 hrs/week network-wide.

Timeline

Build: 9 weeks. Deploy: 3 weeks per wave. Optimization: 10 weeks protocol alignment.