Case Study: $10B Credit Union Improves Account Takeover Fraud Detection by 56%
In a proof of value, a $10B credit union used TAZI's account takeover fraud detection to improve true positives by 56% and detect $2M in additional fraud over two months.
TAZI Team
The client
A $10B credit union based in New England with about 1 million members.
The challenge
The credit union faced ongoing account takeover (ATO) fraud. Its internal Python-based AI model became unserviceable after the developer who built it left. The model also had a 50% false positive rate amid high transaction volumes, so the fraud team spent time investigating legitimate transactions. The credit union needed a solution that was more scalable and transparent than its black-box model.
The approach
In a proof of value, TAZI worked with the credit union's data science team to integrate TAZI's ATO fraud detection solution through a REST API.
Results
- 56% improvement in true positives
- $2M in additional fraud detected across two months
- 15% of fraud losses prevented
- 53% efficiency gain