Project Info
NexaQuanta’s AI transformation solution for a FinTech client leverages Generative AI and Variational Auto-Encoders to detect fraud across multi-blockchain, multi-currency transactions. By modelling user behaviour and identifying anomalies, the solution significantly reduces false positives and scales seamlessly with growing transaction volumes.
- Client : A Fintech Solution Provider
- Category : AI Transformation of Financial Services, AI Transformation of Software Applications
Project Info
Client & Industry:
- FinTech Software Provider
Challenge:
- Multi-blockchain, multi-currency transaction complexity
- Imbalanced transaction volumes defeated rules-based detection
- Excessive false positives undermined existing monitoring tools
AI Solution:
- Gen AI (Variational Auto-Encoders) model each user’s normal behaviour
- Deviations from individual baseline flagged as potential fraud
- Lightweight AI plugin integrates into existing infrastructure
Outcome/Benefits:
- User-level fraud detection across multiple blockchain networks
- False positive rates reduced vs rule-based approaches
- Scales to growing transaction volumes without degradation
AI Technologies:
- Gen AI time-series model (Variational Auto-Encoders), ML, AWS Cloud, Blockchain
