Healthcare Shorts: MedPolicy AI Assistant

[Windows Server]

Story Background

Large healthcare payers have a complex contact center where their agents need to know lots of detailed information about claims, cases and medical policies. Helping these agents with an AI assistant that is knowledgeable about medical policies will help them assist providers and members as they call in with medical policy questions.

Business Challenge 

Healthcare payors are hiring and training new agents on a regular basis due to turn over or growth. These agents need to assist members and providers which is difficult and time-consuming. The amount of information that they need to know is spread out across disparate systems and much of it is tribal knowledge, which is hard to train on. It is difficult for these agents to find the correct answer and respond to member in a timely fashion. These create poor customer service experience for the member & providers who are calling in, and extreme frustration for the agent as they want to do their job well but do not have the proper training or tools to do it quickly and accurately. 

Business Outcomes 

Using a medical policy AI chat bot will:

• Help the agents provide faster response times,
• First time call resolution
• Better member satisfaction
• Reduce the time for prior authorizations for providers
• Help process claims faster
• Help payors get paid quicker
• Lower call center expenses.

Solution Overview 

This solution starts with using Azure AI studio to implement Azure OpenAI models and embedding, uses as your blob storage to upload documents, uses API‘s to connect to data sources, Azure API management to secure and store API information, uses Azure App services to configure and publish the to a web app for user consumption. 

https://youtu.be/Z7PsA6_mj4k

 

Thanks for reading, Shelly Avery |Email, LinkedIn 

If you are interested in more content like this, then follow https://aka.ms/HealthcareShorts 

Please follow aka.ms/HLSBlog for all this great content.

View the full article