Sun Finance automates ID extraction and fraud detection with generative AI on AWS
Amazon AWS AI Blog / 5/1/2026
💬 OpinionDeveloper Stack & InfrastructureTools & Practical UsageModels & Research
Key Points
- Sun Finance built an AI-powered identity verification pipeline by combining Amazon Bedrock (LLM structuring), Amazon Textract (OCR/field extraction), and Amazon Rekognition (image understanding) on AWS.
- The approach significantly improved ID extraction accuracy from 79.7% to 90.8%, while cutting per-document costs by 91% and reducing processing time from as much as 20 hours to under 5 seconds.
- Using a hybrid design that pairs specialized OCR extraction with LLM-based structuring outperformed workflows that relied on either Textract or the LLM alone.
- The post also outlines an AWS serverless fraud detection architecture that uses vector similarity search to identify potentially fraudulent identities or documents.
In this post, we show how Sun Finance used Amazon Bedrock, Amazon Textract, and Amazon Rekognition to build an AI-powered identity verification (IDV) pipeline. The solution improved extraction accuracy from 79.7% to 90.8%, cut per-document costs by 91%, and reduced processing time from up to 20 hours to under 5 seconds. You'll learn how combining specialized OCR with large language model (LLM) structuring outperformed using either tool alone. You'll also learn how to architect a serverless fraud detection system using vector similarity search.
Related Articles
Black Hat USA
AI Business
The foundational UK sovereign-AI patents are filed. The collaboration door is open.
Dev.to
Building a Shopify app with Claude Code — spec-driven development and pricing design
Dev.to
The AI Habit That Pays Dividends (And Takes Zero Extra Time)
Dev.to
From Chaos to Clarity: AI-Powered Client Portals for Designers
Dev.to
