What challenges did Financial services institutions face before AI implementation?

Traditional customer onboarding in banking is manual, slow, error-prone, and costly, relying on physical document processing and in-person verification. Limited service hours and accessibility issues affect customer experience and operational efficiency. There was a need for a secure, automated solution providing 24/7 onboarding that improves security and reduces operational costs.

What was the impact of the AI solution?

Enabled fully digital onboarding without paper forms or branch visits. Automated KYC verification and identity checks, reducing manual workloads and operational costs. Customers are onboarded securely within minutes, improving accessibility and overall experience. Provided personalized, 24/7 virtual assistant support.

Source

ProductionEvidence: Low40/100

Automated User Onboarding for Financial Services with a Digital Assistant Powered by Amazon Bedrock

Use case typeCustomer service automationUpdated Jun 13, 2026

Financial services institutions enhanced customer onboarding by deploying Penny, a digital assistant powered by Amazon Bedrock and a custom LangChain agent. The assistant automates KYC document processing, identity verification using Amazon Textract and Amazon Rekognition, and conversational interactions for fully digital onboarding without branch visits. The system uses Retrieval-Augmented Generation (RAG) and orchestrates multiple AWS services via API calls to deliver 24/7 personalized onboarding support. Users receive email confirmation upon secure, rapid account creation, significantly reducing manual efforts and operational costs.

Organization: Financial services institutions
Industry: Finance
Location: United States
Published: October 2024

Reported outcomes

Strategic outcomes

New product / capabilityEnabled fully digital onboardingRisk & complianceAutomated KYC and identity checksCustomer experience & trustImproved secure onboarding experienceSpeed & agilityProvided 24/7 virtual assistant support

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

Customer: Financial services institutions
Provider: AWS
Maturity: Production
Linked source: AWS Machine Learning Blog

Users receive email confirmation upon secure, rapid account creation, significantly reducing manual efforts and operational costs

Customer identity supportedSource describes one deploymentMaturity supported

Primary read

Use case focus

Showing 3 of 3

1AI-powered Digital Assistant
2Customer Onboarding Automation
3Conversational AI

Traditional customer onboarding in banking is manual, slow, error-prone, and costly, relying on physical document processing and in-person verification.
Limited service hours and accessibility issues affect customer experience and operational efficiency.
There was a need for a secure, automated solution providing 24/7 onboarding that improves security and reduces operational costs.

Built Penny, an AI-powered digital assistant using Amazon Bedrock foundation models (Anthropic's Claude 3.5 Sonnet) and LangChain agents for workflow orchestration.
Used AWS AI services including Amazon Textract for document text extraction, Amazon Rekognition for face verification, AWS Amplify for frontend hosting, Amazon ECS Fargate for container deployment, Amazon DynamoDB for database, and others.
Implemented RAG for context-aware responses and prompt engineering to ensure seamless, accurate onboarding conversations.
Integrated email confirmation via Amazon Simple Email Service (SES).

Technologies

Amazon Bedrock Amazon Textract Amazon Rekognition Amazon S3 Amazon API Gateway Amazon DynamoDB Amazon Simple Email Service AWS Amplify Amazon ECS Fargate Amazon Kendra LangChain

Enabled fully digital onboarding without paper forms or branch visits.
Automated KYC verification and identity checks, reducing manual workloads and operational costs.
Customers are onboarded securely within minutes, improving accessibility and overall experience.
Provided personalized, 24/7 virtual assistant support.

Architecture

The solution includes a web frontend hosted on AWS Amplify, a containerized AI assistant deployed on Amazon ECS Fargate, AWS API Gateway for service APIs, Amazon Textract and Rekognition for document and face verification, Amazon DynamoDB for data storage, and Amazon SES for email notifications. The assistant orchestrates workflows using a custom LangChain agent with prompt engineering and uses Retrieval Augmented Generation (RAG).

Sources & evidence1

Evidence: Low40/100Evidence strength