Inspiration
Most cardholders pay annual fees for premium credit cards but never use the insurance and protection benefits included with them because they're buried in lengthy terms and conditions. Cover Catch uses AI to bridge that awareness gap, ensuring users automatically discover and claim benefits they're already entitled to instead of leaving value unclaimed.
What it does
Cover Catch monitors a card member's transaction stream in real time and automatically detects purchases that qualify for card-linked protections, including Purchase Protection, Return Protection, and Travel Delay Insurance.
When a qualifying event is detected—such as a damaged electronics purchase, an eligible return, or a delayed flight—the system identifies the appropriate benefit, matches the transaction to the correct protection, and automatically pre-fills the claim form. Instead of searching through benefit guides and manually entering claim details, the cardholder only needs to review and submit the claim.
How we built it
Real-Time Transaction Monitoring
- Designed an algorithm to monitor transaction streams in real time.
- Processes incoming transactions to identify purchases eligible for card protections.
- Supports scalable event-driven processing for future production integrations.
Detection & Benefit Matching Engine
- Implemented a rules-first classification engine based on:
- Merchant category
- Purchase amount
- Transaction timing
- Benefit eligibility rules
- Automatically matches qualifying purchases to the correct benefit type:
- Purchase Protection
- Return Protection
- Travel Delay Insurance
- Built to support future ML-based confidence scoring for ambiguous cases.
Automated Claim Pre-fill
- Extracts transaction details and automatically populates:
- Purchase date
- Merchant name
- Transaction amount
- Benefit type
- Enables customers to review and submit claims instead of manually filling lengthy forms.
Customer Dashboard
- Developed an intuitive interface that displays:
- Real-time detected benefits
- Eligible claims
- One-click claim initiation
- Claim progress and status
- Provides a seamless experience from detection to submission.
Backend Services
- Backend APIs manage:
- Benefit entitlement verification
- Claim creation
- Claim submission
- Approval workflow
- Claim status tracking
Tech Stack
| Layer | Technologies |
|---|---|
| Frontend | React, Vue.js |
| Backend & APIs | Node.js, Spring Boot |
| Transaction Processing | Python, Apache Kafka |
| Rules & ML Engine | scikit-learn, TensorFlow |
| Cloud Platforms | AWS, Google Cloud Platform (GCP) |
| Databases | MySQL, DynamoDB |
Challenges we ran into
- Mapping ambiguous merchant categories to the correct insurance benefit without generating excessive false positives.
- Designing a rules engine that remains transparent and explainable instead of becoming a black-box model.
- Building a realistic transaction simulation that closely resembles real customer spending patterns within the limited hackathon timeframe.
- Optimising detection accuracy while maintaining fast real-time processing.
Accomplishments that we're proud of
- Built a complete end-to-end workflow from raw transaction → benefit detection → pre-filled claim.
- Successfully designed a real-time transaction monitoring and benefit detection engine.
- Automated claim pre-filling to significantly reduce customer effort.
- Created a modular architecture that is ready for integration with real issuer transaction APIs.
- Designed backend workflows to support benefit verification, claim submission, and approval management.
What we learned
Building AI for financial assistance requires more than accurate predictions—it demands transparency, explainability, and trust. We also learned that combining deterministic business rules with machine learning creates a practical foundation for financial products where reliability is critical.
What's next for Cover Catch
- Integrate directly with issuer transaction APIs instead of synthetic transaction feeds.
- Introduce ML-based confidence scoring to reduce false positives using production transaction data.
- Expand support beyond Purchase Protection, Return Protection, and Travel Delay Insurance to cover the full suite of credit card protections.
- Add proactive claim reminders before eligibility windows expire.
- Continuously test and optimise the detection engine to improve:
- Detection accuracy
- Claim pre-fill quality
- Reduction in unclaimed card benefits
Resources
Google Cloud Pub/Sub Documentation
https://docs.cloud.google.com/pubsub/docsStripe Issuing API Documentation
https://docs.stripe.com/issuingAWS Lambda Documentation
https://docs.aws.amazon.com/lambda/
Log in or sign up for Devpost to join the conversation.