Case Study: UCSF Health achieves automation of 14M annual e‑faxes and saves 30,000 hours with H2O.ai Document AI

Applying H2O Document AI to Automate Workflows in Healthcare

UCSF Health, a leading academic medical center, struggled with more than 14 million incoming e-faxes per year that required manual review to identify document type, extract patient and referral data, and route workflows—work equivalent to over 100 full-time employees and causing delays in patient access. To address this, UCSF partnered with H2O.ai and adopted H2O Document AI as part of an automation initiative to reduce the heavy burden of semi-structured fax processing.

UCSF built an Intake Automation pipeline that digitizes faxes (RightFax), stores PDFs, and calls the H2O Document AI API to classify pages, extract fields, and populate metadata for streamlined human review and EHR entry. Powered by H2O.ai, the solution is projected to save more than 25,000 hours (~12 FTE) of access center staff time and 5,000 hours (~2.5 FTE) of clinician time, improving patient experience and enabling UCSF Health to handle growing referral volumes without adding administrative headcount.

UCSF Health

Bob Rogers

Expert in Residence for AI