Look Up Claim Policies During Claim Settlement Using Generative AI
This demo will show you how you can use generative AI as a claim policy advisor to help analysts navigate policies and procedures during claim settlement. This feature provides the ability for organizations to ingest claim policies and procedures into an LLM, a Large Language Model, and then use natural language to ask a generative AI chatbot questions about those policies during claim settlement.
Every search engine today provides search results from available web content. The problem is that enterprises have proprietary information that they want in a secure LLM. This feature provides the ability to have a secure LLM learn your claim policies and then put that knowledge to use right where they need it, improving the efficiency of claim analysts. The business outcome is that it improves adherence to claim policies and procedures and reduces the onboarding time of new claim analysts.
I've been recently hired as a claims analyst. Here I am at the Claims tab with a couple of claims I need to investigate. So let's start with the first one. This is a deduction, and I want to check that I'm doing it properly. And luckily, I've got an AI Claim Advisor at my disposal, and it supports one or many topics. I'm going to ask it a very generic question. I'm going to ask it, "How do I handle a short pay?"
So with the configuration of the AI Advisor, our business policy documents were uploaded to each topic in the AI agent. So it's a LLM or a large language model that also knows our claim policy rules. So here, I'm able to get an answer without going through all the documents just by asking.
So let's say it's telling me that I should always start with a debit memo because it documents the reasons for the short pay, and it's going to give me a place to start my investigation. You can also see there are sources provided for each answer, and I can drill into them to see the actual source in the context. So here is that answer in its complete context.
Thank you for watching. You can refer to the Channel Revenue Management 25A, What's New document for more information on this feature.