moonbatant.
All work

Automating fax-based referral intake with an AI agent

Innovaccer · Product Manager · 2022–2025

4x
Per-user operational efficiency
40%
Faster turnaround time
AI agent Healthcare ops Automation Workflow

📰 Featured in Fierce HealthcareOlympia Orthopaedic Associates rolls out Innovaccer’s AI agents for referrals.

Context

In US healthcare, specialty referrals still move over fax. For specialty clinics, a human reads each inbound fax, extracts the patient and referral details, and keys them into the system — slow, error-prone, and a poor use of clinical-ops time. At Innovaccer I owned the product work to automate this.

The problem

Manual intake was the bottleneck. Turnaround time (how long from fax arriving to referral actioned) was long, throughput was capped by headcount, and accuracy depended on whoever was reading the page that day.

Approach

I shipped an AI intake agent that reads inbound faxes, extracts the structured referral data, and routes it into the workflow — turning a manual transcription step into a reviewed, automated one, with a human confirmation step kept in place where a mis-read would be costly. The referral agents were rolled out with real specialty-clinic customers, including Olympia Orthopaedic Associates, and the launch was covered by Fierce Healthcare.

Impact

  • 4x improvement in per-user operational efficiency.
  • 40% faster turnaround time on referral intake.

Reflection & tradeoffs

This was an early, concrete example of the pattern I’ve kept building since: automate the high-volume, low-judgment transcription work, but keep a human review step where being wrong is expensive.