ePortal Referral Intake with Intelligent Bots and AI Agents
Project Insight
Super Charge ePortal Referral Intake
About This Project
Managing referrals from over 30 ePortals across multiple states presented a significant challenge for a large home health and hospice organization. The complexity increased with the need for automation to handle referral acceptance, decline, or other decisions based on branch, payer, and coverage area relationships. Our mission was clear yet demanding:
System Design: Create a seamless system to automate the intake of diverse referrals from various ePortals such as AIDA, AIDIN, CuraSpan, EPIC, ExtendedCare, and others.
Referral decision: Automate referral decision based on specific branch, coverage and payor contracts. Trigger operations in real time to intervene if required and reinitiate automation after the manual step at will.
Centralized Dashboard: Enable the referral intake team across multiple regions and states to view referral activities through a smart, centralized, and lean dashboard.
Continuous Optimization: Provide tools for continuous process fine-tuning and generate reports with new insights to help leadership adjust strategic and tactical plans periodically.
Our Solution
We deployed a team of bots and AI agents trained to mimic human patterns and decisions. These intelligent bots were responsible for:
Referral Gathering: Collecting referrals from numerous portals and funneling them through a central dashboard with access control at the branch, region, and central levels.
Standardization: Beyond gathering referrals, the bots standardized the process, definitions, and reporting to ensure consistency and accuracy.
Results
The implementation of the bots and AI agents led to significant improvements:
Efficiency and Scalability: Manual efforts in referral processing were drastically reduced, and the system provided enormous scalability.
FTE Reduction: The centralization and automation reduced the Full-Time Equivalent (FTE) count by more than 50%.
Speed: A decision on a referral is now made within 20 seconds, including the downloading and assessment of available patient documentation.
Domain & Technology
- Microsoft Power Automate
- Azure SQL
- Azure BLOB Storage
- OpenAI’s GPT
- Langchain
Explore more of our work