AIT Prevention

Companies in virtually every sector can be affected by AIT. Fraudsters target apps and platforms that have high volumes of international users and use SMS for 2FA and OTP. From social media and travel apps to major banks, tech giants, and e-commerce sites, this fraud can affect nearly any large user-based business. 

The rewards are in the SMS delivery chain. SMS aggregators or other entities in the delivery chain benefit by charging to carry and terminate generated SMS traffic. Sometimes the traffic is generated using fake numbers, and therefore cannot be delivered, however another common practice is to use valid (real) numbers, and then “trash” the messages before they ever reach the recipient. The bad actor profits by charging upstream for messages that they never deliver downstream. 

CAPTCHAs and other validation techniques such as 2FA emails add unnecessary steps to the user experience, and are increasingly defeated as fraudsters improve their techniques. The incentives for SMS traffic inflation are in SMS and only by “following the money” can you really detect fraudulent activity effectively. 

The business model of SMS platforms and SMS aggregators is based on total traffic volumes. To put it simply: they lack incentives to detect and block SMS generation of any kind. While they may offer fraud detection, it is difficult to prove the efficacy of a carrier-based solution or the validity of the entire message chain. Carrier-neutral solutions remove the inverse system of incentives that fuel AIT in the first place.

Yes. There is no need to access the content of SMS messages in order to detect fraudulent patterns. By monitoring and analyzing the behavior patterns of the messages, AI shield is able to fingerprint normal traffic fluctuations and identify artificially generated SMS. When configuring the API query, message parameters can be fully anonymized for analysis. The solution can be integrated on-prem or in a compliant or local cloud environment to comply with data privacy standards.

The latency added to any message delivery is milliseconds and is completely unnoticeable by the end user. Deliverability of legitimate messages is not affected by the API query. 

Absolutely. Offering fraud prevention as a service to enterprise customers creates value for app analytics, CPaaS platforms, and more. 

Fraudsters frequently use real numbers to generate fake accounts, so using standing allow lists would effectively defeat the security. Complaints about non-delivery of OTP and 2FA from legitimate users is very unlikely. 

Yes and no. We see that in most cases, fraudsters shift their efforts to easier targets, but leave a smaller probe campaign running. If this probe attack becomes successful, they return to a full-scale attack. It is important to maintain the deterrent in place to discourage the return of these attacks.