Generative AI-powered Financial Fraud Detection System
FraudGuard AI is a web-based prototype developed for the AMAD Hackathon to help financial institutions detect suspicious transactions using Generative AI.
The system analyzes transaction data, customer behavior, device information, and risk indicators using Gemini API, then generates a clear explanation of the detected risk, recommended actions, and automatically updates reports, alerts, fraud cases, notifications, and the dashboard.
🌍 Website FraudGuard AI
🎥 Demo VideoWatch Demo Video
- AI-powered transaction analysis
- Fraud risk scoring
- AI-generated fraud explanation
- Smart recommendations
- Automatic reports
- Fraud case management
- Alerts & notifications
- Interactive dashboard
- Arabic & English interface
- Google AI Studio
- Gemini API
- React
- TypeScript
- Express
- Tailwind CSS
- CSV Dataset
- Prompt Engineering
The project uses a synthetic CSV dataset containing 500 realistic financial transactions representing normal, suspicious, and high-risk scenarios while preserving customer privacy.
To assist financial institutions in detecting suspicious transactions faster by combining Generative AI with an interactive banking dashboard that improves transparency and supports decision-making.
✅ MVP Completed
Current version includes:
- Transaction Analysis
- Dashboard
- Reports
- Alerts
- Fraud Cases
- Notifications
- AI Assistant
- Arabic & English Support
Lama Raied Alghamdi
Individual Project – Team Marsad
This project was developed for educational and hackathon purposes.
