AI-driven Insurtech
REVOLUTIONIZING VEHICLE DAMAGE ASSESSMENT
REVOLUTIONIZING VEHICLE DAMAGE ASSESSMENT
Client Overview
Our client is an AI-powered automotive insurtech platform revolutionizing vehicle damage assessment and insurance claims. As the first system of its kind, it leverages advanced AI dent detection technology to automatically identify hail damage, count dents, assess repair needs, and flag total vehicle losses. Trusted by leading insurers and repair networks, the platform streamlines automotive insurance claim workflows by delivering real-time damage reports and actionable insights, enhancing efficiency and accuracy in vehicle damage inspections.
The Challenge
After years of development, the AI-powered automotive insurtech platform was capable of automatically detecting hail damage with unmatched precision. Before a full rollout, it needed to be tested on real vehicles using everyday mobile phones in real-world conditions. This pilot aimed to evaluate the entire automated vehicle damage assessment workflow - from capturing high-quality mobile video footage and uploading media to running AI dent detection inference and generating detailed reports within minutes.
The key challenge was achieving accurate hail damage detection and dent counting while ensuring speed, reliability, and consistent performance across hundreds of submissions. Additional expectations included supporting consistent media quality across Android and iOS devices, completing inspection and reporting in under 10 minutes per vehicle, and handling a high volume of damage reports with minimal failure or manual intervention.
Our Solution
A lightweight web application was designed to test the AI-powered hail damage detection workflow in real-world conditions. The app enables seamless mobile video capture and triggers a fully automated AI dent detection and damage assessment pipeline, ensuring a quick, reliable, and user-friendly insurance claim process.
Video-First Data Collection
Optimized for mobile, users were guided to record short, portrait-mode videos of key vehicle areas, including hood, roof (left and right), and both sides, capturing high-quality footage tailored for AI dent detection algorithms. Optional photos provided additional context for more accurate vehicle damage assessment.
Built-In Guidance for Consistency
Visual instructions and filming examples were embedded directly into the app to ensure consistent media quality and reduce variability. This guidance made it easy for field agents and everyday users to capture usable footage across multiple platforms, including Android and iOS devices, improving the reliability of insurance claim automation.
Instant Uploads and Background Processing
Uploads were triggered automatically after footage capture. The system leveraged AWS Step Functions to run a fully automated backend workflow that included video reformatting, AI dent detection via Roboflow, and damage inference - all without user intervention. This seamless process facilitated fast and reliable vehicle damage detection.
Automated Report Generation
The AI processing results were compiled into a simple PDF report detailing dent counts and damage insights. Reports were securely stored in AWS S3 and made available to users within minutes, enabling prompt real-time vehicle damage reporting and faster claim decisions.
Designed for Testing at Scale
Every step - from vehicle ID and upload time to report status and weather conditions - was logged, allowing monitoring of over 600 vehicles during the pilot. Multi-device support and hardware rotation ensured cross-platform compatibility. This comprehensive AI-driven solution validated the platform’s scalability, accuracy, and operational efficiency in automotive insurance claims automation.
Tech Stack
Next.js
Nest.js
PostgreSQL
AWS EC2
AWS Lambda
AWS SES
Python
React Native
Results
The pilot demonstrated the viability of a fully digital, AI-powered hail damage assessment process, delivering fast and accurate vehicle damage reports from mobile video upload to automated PDF generation.
- Over 500 vehicles were successfully processed using everyday mobile devices in real-world conditions, validating the platform’s scalability.
- More than 80% of damage reports were generated in under 10 minutes, meeting the benchmark for real-time insurance claim assessments.
- Cross-platform compatibility was confirmed by handling videos from both iOS and Android devices, ensuring consistent media quality standards.
- Optional photo capture and 3D media review tools increased submission confidence and enhanced clarity for insurers.
- Automated dent counting and report generation eliminated manual evaluation bottlenecks, enabling scalable and efficient claim processing with consistent accuracy.
The client leveraged the collected data to fine-tune the AI dent detection models and optimize video upload workflows. Early operational risks such as app crashes and upload failures were identified and mitigated with fallback mechanisms and improved logging.
What once took days now happens within minutes, pushing automotive insurance claims closer to fast, touchless, and AI-driven vehicle inspections.
5x
faster damage report generation
0
manual dent counting needed
100%
automated end-to-end workflow