Human Ingenuity meets AI- Intelligent Agent Framework to Transform the Healthcare Supply Chain

Since introducing ChatGPT and other AI technologies in 2024, the healthcare supply chain has experienced excitement and skepticism. One group sees immense opportunities in this technology, while the other fears it may be a fleeting bubble. Both perspectives have merit. The potential to generate instant answers to complex questions is revolutionary, but using AI ethically and responsibly is essential.

Since its inception, SupplyCopia has invested in ChatGPT and other large language models (LLMs). The ability to analyze vast amounts of structured and unstructured data is promising, but dealing with AI "hallucinations"—inaccurate or nonsensical outputs—has been challenging. The healthcare supply chain is particularly complex due to several factors:

1. Thousands of suppliers offer billions of products and services to health systems worldwide, each with unique methods.
2. Intricate relationships exist between manufacturers, product distributors, health systems, and group purchasing organizations.
3. Data asymmetry exists in basic product information, complicating comparisons.
4. Each region has distinct rules, compliance mechanisms, and regulations.
5. Competitive interests often hinder collaboration within the ecosystem.
6. Strict HIPAA compliance, data security, and privacy requirements must be met.

Initial results from our AI integrations were promising but marred by hallucinations, producing irrelevant or incorrect answers. The turning point came when we combined LLMs with human intelligence through prompt engineering. This hybrid approach resulted in more accurate and insightful responses. We realized that LLMs alone are insufficient in the healthcare supply chain, necessitating the integration of human expertise.

This led to the creation of our unique "hybrid control tower" architecture. This system integrates a data lake containing health system data into SupplyCopia applications. Our intelligent agent, "Ask the BEE," operates on this framework to answer user questions instantly. Whether you're a CFO, finance manager, clinician, supply chain leader, or value analysis professional, you can ask the BEE and receive immediate, actionable insights. Users can also generate visualizations and logic from the LLMs, significantly reducing the time needed for analysis.

We also faced another challenge: user training. We started with the premise that it is as simple as asking a question and getting an answer. Similar to when we went from writing long emails to sending a text! We had a surprise waiting for us. When we gave users access to the BEE, the users simply didn’t know how or what questions could be asked! Years of using rigid dashboards had conditioned their mind to look at data and analysis in a certain way. We addressed this issue with a question bank; we started with the most frequently asked question and the ability for each user to customize their question bank.

We have now deployed the intelligent agent in all of our clients, and the results have been phenomenal. The users are finally escaping from the dashboard deluge. The control is back where it belongs, i.e., stakeholders. Now, they have the power to create analysis and insights on demand.

Our current applications of the ChatGPT 4o have been around the analysis and a chatbot for customer support; we look forward to applying the framework to as diverse areas as Accounts payable automation, revenue leakage, procedure card optimization, procedure scheduling, contract compliance, and medical device management.

With the SupplyCopia intelligent agent, tasks that once took weeks or months can now be accomplished in minutes. We invite you to join this revolution and simplify the healthcare supply chain. Explore our white papers written by our lead data scientists for a deeper understanding. If you have any questions, interact with our ChatBot on the website or email us at [email protected], and we will respond promptly.

?

?

?

?

?