Agentic Artificial Intelligence as a Catalyst for Administrative Modernization: The Beginning of the End for Traditional Fax Workflows in Healthcare

Read the full article See related articles

Discuss this preprint

Start a discussion What are Sciety discussions?

Listed in

This article is not in any list yet, why not save it to one of your lists.
Log in to save this article

Abstract

Background

Healthcare has witnessed administrative staffing roles balloon to twice the number of employed clinicians, resulting in $950 billion per year in administrative costs to deliver healthcare. Administrative workflows, like fax routing, are ripe for automation given the high human labor cost necessary to complete these tasks. Fax transmissions remain a key mode of communication in modern healthcare, requiring substantial manpower, and incurring significant, though not well-characterized, costs to health systems. Opportunities may exist for agentic artificial intelligence (AI) to automate this administrative task.

Methods

This quality improvement study was performed in 2 phases at Duke University’s Division of Cardiology, a single tertiary-care cardiac referral center. The first retrospective phase employed an observational time study design surveying manual fax routing processes at 3 representative cardiology clinics from April 1, 2024, to July 12, 2024. The second phase quantified all inbound faxes received through the division’s communication hub from July 1, 2025, to December 31, 2025, and applied direct labor costs observed in the time study to quantify the economic burden of manual fax routing across the hub.

Results

The observational time study (Phase 1) demonstrated fax routing processing times ranging from 4.4 to 9.4 minutes depending on fax type, with a mean processing time of 6.0 minutes per fax. On average, the ambulatory clinics received 1,694 faxes per month and spent 169.1 person-hours routing faxes. The divisional communication hub (Phase 2) received 24,420 faxes over the study period, averaging 4,070 inbound faxes and 13,341 pages of information per month. Extrapolating direct labor efforts from the time study, 407 person-hours per month were spent processing inbound faxes. For our institution, this translated to $10,663.40 in total monthly costs, roughly 2.5 full-time equivalents.

Conclusion

Manual fax routing represents a substantial, measurable, and previously under-characterized operational and administrative burden. Given the significant opportunity to reduce labor time and costs, our study establishes fax routing as a high-value target for automation. Future work is needed to determine the impact of AI-automated fax routing on the time, labor, accuracy, and economics within clinical settings.

Article activity feed