What challenges did EXL face before AI implementation?

Manual insurance underwriting was time-consuming, requiring days of reviewing hundreds of pages of documents by underwriters. Concerns about generative AI hallucinations posed risks to reliability, and regulatory compliance with India's PII regulations was critical. EXL needed a scalable, secure, and cost-effective solution to automate underwriting while maintaining accuracy and data privacy.

What was the impact of the AI solution?

The AI assistant reduced underwriting processing time from several days to a few hours. Underwriting costs dropped by up to 80%, significantly improving insurer efficiency. The enhanced accuracy and privacy compliance boosted customer trust and accelerated market adoption. The rapid 60-day development cycle provided EXL competitive advantages and paved the way for further generative AI applications on AWS.

Source

ProductionEvidence: Medium65/100

EXL Transforms Insurance Underwriting with Generative AI Assistant Built on Amazon Bedrock

Use case typeDocument automationUpdated Jun 13, 2026

EXL, a global data and AI company, sought to reduce the manual effort in insurance underwriting by developing a generative AI virtual assistant on AWS to streamline document processing and evaluation. The company developed LDS™ Underwriting Assist, a generative AI chatbot on Amazon Bedrock integrating multiple large language models and AWS services including Amazon Textract, Amazon Comprehend, and Amazon Kendra, ensuring data privacy and compliance with India's PII regulations. Launched in August 2024 after 60 days of development, the solution reduced underwriting time from days to hours and cut costs by up to 80%. The platform improved accuracy, data security, and accelerated customer adoption with positive market impact.

Organization: EXL
Industry: Insurance
Location: India
Published: April 2026

Reported outcomes

−80%

costCost savings

Strategic outcomes

Speed & agilityReduced underwriting processing timeCost efficiencyLowered underwriting costsCustomer experience & trustImproved customer trust and adoptionCompetitive differentiationCreated competitive advantages

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

Normalized claim

Cost: 80% decrease

Amazon AWS Customer StoriesApr 29, 2026Customer storyInferred claimMedium evidence strength

Underwriting costs dropped by up to 80%, significantly improving insurer efficiency.

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

Customer: EXL
Provider: AWS
Maturity: Production
Linked source: Amazon AWS Customer Stories

AWS's managed cloud environment ensured security, compliance, and operational efficiency without infrastructure management overhead

Customer identity supportedSource describes one deploymentMaturity supported

Primary read

Use case focus

Showing 3 of 3

1Generative AI Assistant
2Document Processing Automation
3Compliance Automation

Manual insurance underwriting was time-consuming, requiring days of reviewing hundreds of pages of documents by underwriters.
Concerns about generative AI hallucinations posed risks to reliability, and regulatory compliance with India's PII regulations was critical.
EXL needed a scalable, secure, and cost-effective solution to automate underwriting while maintaining accuracy and data privacy.

EXL partnered with AWS to build LDS™ Underwriting Assist using Amazon Bedrock, which provides access to multiple LLMs for model testing and fine-tuning to reduce hallucinations.
The solution incorporated Amazon Textract for document data extraction, Amazon Comprehend for text analysis and PII redaction, and Amazon Kendra for enhanced search capabilities.
A custom API built with FastAPI and containerization simplified deployment and integration.
AWS's managed cloud environment ensured security, compliance, and operational efficiency without infrastructure management overhead.

Technologies

Amazon Bedrock Amazon Textract Amazon Comprehend Amazon Kendra Sonnet 3 LLM FastAPI

The AI assistant reduced underwriting processing time from several days to a few hours.
Underwriting costs dropped by up to 80%, significantly improving insurer efficiency.
The enhanced accuracy and privacy compliance boosted customer trust and accelerated market adoption.
The rapid 60-day development cycle provided EXL competitive advantages and paved the way for further generative AI applications on AWS.

Architecture

Architecture uses Amazon Bedrock with multiple LLMs, Amazon Textract for document data extraction and PII redaction, Amazon Comprehend for text analysis, Amazon Kendra for search, FastAPI for interfacing, and containerization in a secure AWS environment.

Sources & evidence1

Evidence: Medium65/100Evidence strength