How InsuranceDekho transformed insurance agent interactions using Amazon Bedrock and generative AI

This article discusses how InsuranceDekho, an Indian insurance aggregator, used Amazon Bedrock and generative AI to create a chatbot assistant that can accurately answer customer queries about insurance policies. The chatbot allows insurance agents to quickly get information without needing to consult subject matter experts, improving response times and customer experience.

Specifically, the article covers:

• The challenge of long response times when agents had to consult experts for policy details
• Using Amazon Bedrock and a retrieval-augmented generation (RAG) approach to create the chatbot
• Key components like language models, embedding models, vector databases, and caching
• The implementation workflow involving ingesting policy docs, embedding, vector search, and generation
• Results showing an 80% reduction in response times after deploying the chatbot
• Conclusion on how the solution empowered agents and improved customer service