Customer support · Production

InsuranceDekho cuts customer query response time by 80% with Amazon Bedrock and RAG

The problem

InsuranceDekho's insurance advisors, especially newer ones, had to consult subject matter experts for policy-specific queries, introducing delays of a few minutes before they could respond to customers and creating a bottleneck that led to lost prospects and added costs.

Workflow diagram · grounded in source

Policy document ingestion

integration

“These documents are processed by the embedding model, which converts the textual content into high-dimensional vector representations, capturing the semantic meaning of the text. After the embedding model generates the vector representat…”

Advisor query via chatbot

trigger

“our chatbot serves as the entry point, facilitating seamless interaction between the insurance advisors and the underlying response generation system”

Cache lookup

routing

“This solution incorporates a caching mechanism that uses semantic search to check if a query has been recently processed and answered. If a match is found in the cache (Redis), the chat assistant retrieves and returns the corresponding r…”

Intent classification

ai_action

“the query goes to the intent classifier powered by Anthropic's Claude Haiku. It analyzes the query to understand the user's intent and classify it accordingly. This enables dynamic prompting and tailored processing based on the query typ…”

RAG retrieval from policy docs

ai_action

“the intent classifier passes the query to the retrieval step, where a semantic search is performed on a vector database containing insurance policy documents to find the most relevant information, that is, the context based on the query”

Response generation

output

“the retrieved context is integrated with the query and prompt, and this augmented information is fed into the generation process. In the final generation step, the actual response to the query is produced based on the external knowledge …”

Reported outcome

InsuranceDekho's RAG chat assistant enabled insurance advisors to address customer queries autonomously without SME involvement, achieving an 80% decrease in response time and improving sales, cross-selling, and overall customer service.

Reported metrics

Customer query response time80%

Sales and cross-sellingimproved sales, cross-selling, and overall customer service experience

SME reliance for advisor querieswithout the constant need for SME involvement

Reported stack

Amazon BedrockAnthropic's Claude HaikuRAGAWS PrivateLinkembedding model

Source

https://aws.amazon.com/blogs/machine-learning/how-insurancedekho-transformed-insurance-agent-interactions-using-amazon-bedrock-and-generative-ai?tag=soumet-20

Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

InsuranceDekho's RAG chat assistant enabled insurance advisors to address customer queries autonomously without SME involvement, achieving an 80% decrease in response time and improving sales, cross-selling, and overa…

What tools did this team use?

Amazon Bedrock, Anthropic's Claude Haiku, RAG, AWS PrivateLink, embedding model.

What results were reported?

Customer query response time: 80%; Sales and cross-selling: improved sales, cross-selling, and overall customer service experience; SME reliance for advisor queries: without the constant need for SME involvement (source-reported, not independently verified).

How is this customer support AI workflow structured?

Policy document ingestion → Advisor query via chatbot → Cache lookup → Intent classification → RAG retrieval from policy docs → Response generation.