How does
Careem
Use AI?
Reduces fraud while minimizing customer disruption
Project Overview
Reducing payment fraud by predicting when to apply temporary payment holds.
Layman's Explanation
Careem’s AI figures out when a payment might be sketchy and puts a short pause on it—without bothering most good customers.
Details
Careem developed a machine learning model to intelligently apply temporary holds on transactions that are likely to be fraudulent. Instead of using fixed rules that can frustrate legitimate users, the model classifies transactions based on behavioral and contextual signals to decide when a hold is warranted. This reduces the number of false positives, improves fraud detection, and preserves a smooth experience for the majority of users. The system continuously learns and updates based on new fraud patterns and customer behavior.
Analogy
It’s like a bouncer who steps in only when someone looks suspicious, not every time someone enters the door.
Machine Learning Techniques Used
More Use Cases in
Transportation & Logistics
Novelty Justification
ML-based fraud holds are now a best practice in fintech, balancing security and user experience.
Project Estimates
