Detecting fraud on international calls with AI

Identifying fraud attempts is a relatively complex task. To fight fraud on its international network, Orange uses Khiops, a software solution based on machine learning.

“Orange has developed Khiops, a tool that speeds up information analysis and data preparation.”

Antoine Darbois, Head of Fraud Detection Solutions at Orange Innovation, explains: “Have you ever had a suspicious call, such as someone calling you from abroad, without giving you time to pick up, to encourage you to call back a premium rate number? To combat this type of fraud, Orange teams have implemented a solution based on machine learning. ”

Orange has developed Khiops, a tool that speeds up information analysis and data preparation. The tool allows automatic analysis of the database of calls made; fed by qualified examples, it calculates a probability rate, the scoring. The data analyst focuses on the results with the highest scores and confirms if they are fraudulent behaviors, allowing them to block those phone numbers to stop the fraud.

Every day, from the more than 200 million international call reports it analyzes, the solution generates around 65,000 alerts.

Read also on Hello Future

Attacks on AI: data cleaning becomes a cybersecurity issue


IAlerting: Protecting Customers from Online Piracy


Improving business operations through AI


How Artificial Intelligence is accelerating cybersecurity analysts’ work

Des algorithmes pour gouverner… les algorithmes !

Algorithms to keep other algorithms in line!