Comarch's Loyalty & Marketing portfolio is an advanced set of solutions dedicated to marketing processes and activities, building loyalty, and maximizing engagement.
FREMONT, CA: Comarch, a global software house delivering customer engagement and marketing solutions, announced a new fraud prevention offering specifically designed to provide robust end-to-end monitoring for enrollment fraud.
The AI-powered Enrollment Security Service is trained on historical data to detect fraudulent new registrations or personal data changes. It analyzes each new registration or personal data change using an ensemble of statistical and Machine Learning natural language processing models.
It can detect in real-time if:
Personal data used for registration was synthetically generated.
An unusual or disposable email address is used.
There is a spike in new member registrations, which may indicate that there is a mass enrollment attack happening.
There are repeated parts of personal data used during registration.
"Enrollment fraud is not only dangerous in the case of data protection and privacy but also when it comes to reputation. You want customers to trust your brand … and not see loyalty program accounts being sold on the dark web," said Maciej Tyczynski, Comarch's Head of Data Science. "We built a service that uses historical member data to train machine learning models to detect what genuine member personal data looks like. We're able to score each enrollment in real-time to say whether it looks genuine or if it comes from a suspicious source, in which case we're able to proactively prevent these accounts from being created in the system."
In addition to preventing new fraudulent accounts from being created, the service can also prevent existing accounts from takeovers. All of this requires minimal setup and no rules to configure. It can be deployed and seamlessly integrated with other Comarch or third-party solutions using message brokers or a Web API. The entire solution can be delivered for integration tests within days from sharing historical data for machine model training.