What challenges did InsuranceDekho face before AI implementation?

Insurance advisors frequently consulted subject matter experts for policy questions, causing delays and additional cost. Agents needed accurate, timely answers without memorizing complex insurance plans. Slow responses risked lost prospects and poorer customer experience.

Source

Evidence: Low35/100

InsuranceDekho transformed insurance agent interactions using Amazon Bedrock and generative AI

Use case typeConversational supportUpdated Jun 13, 2026

InsuranceDekho built a generative AI chat assistant for insurance advisors and POSP agents to answer policy coverage and exclusion questions using only company policy documents. The solution uses Retrieval Augmented Generation with Amazon Bedrock and Anthropic Claude Haiku, Amazon OpenSearch Service as a vector database, and Redis on Amazon ElastiCache for caching. An intent classifier routes queries between a lightweight response path and full retrieval workflow.

Organization: InsuranceDekho
Industry: Insurance
Location: India
Published: November 2024

Planned next steps

The source says the organization plans to achieve Response time decrease: −80%.
The source says this outcome is planned: Faster responses to policy questions.

Why do we believe this?Outcome claims, sources, and evidence checks

Normalized claim

Response time decrease: 80% decrease

AWS Machine Learning BlogNov 18, 2024Blog postExplicit claimLow evidence strength

InsuranceDekho has witnessed a remarkable 80% decrease in the response time of the customer queries to understand the plan features, inclusions, and exclusions.

Why do we believe this deployment?Customer identity, provider attribution, maturity, and source checks

Customer: InsuranceDekho
Provider: AWS
Maturity: Unknown
Linked source: AWS Machine Learning Blog

No explicit deployment-stage evidence found.

Customer identity supportedSource describes one deploymentMaturity evidence evaluated

Primary read

Use case focus

Showing 3 of 3

1Conversational AI
2Customer Service
3Knowledge Retrieval

Built a Bedrock-powered RAG assistant grounded only in insurance policy documents.
Added an intent classifier to route simple questions to a lightweight path and more complex queries to retrieval from OpenSearch.
Used Redis caching to avoid recomputation for repeated questions and enable lower latency.

Technologies

Amazon Bedrock Anthropic Claude 3 Haiku Amazon OpenSearch Service Amazon ElastiCache Redis

Architecture

A Bedrock-powered RAG chat assistant combines an intent classifier, Amazon OpenSearch Service vector retrieval over insurance policy documents, and Redis on Amazon ElastiCache caching. Queries are routed either to a lightweight response path or to the full retrieval-and-generation workflow using Anthropic Claude on Amazon Bedrock.

Sources & evidence1

Evidence: Low35/100Evidence strength