Tinkoff has developed a drop monitoring system

2024: Development of a system to identify drops

Tinkoff has developed a system that automatically allows you to identify drops - customers who draw up cards not for their own use, but for the needs of fraudsters in order to withdraw victims' funds from them. The system allows you to identify such customers even before they receive money received by criminal means, as well as reduce the number of drops in the bank by 2 times. The bank announced this on April 26, 2024.

To create the system, a large array of data was analyzed, millions of operations of Tinkoff customers, including those who used the card exclusively for unfair purposes. With the help of the analysis, a "portrait of a drop" was compiled - signs that indicate a high probability that the client issued a card in order to sell or transfer it to fraudsters. Among the factors: re-registration to another phone number, the receipt of small amounts immediately after opening an account for the so-called "warm-up" drop, and much more.

The portrait formed the basis of the work of a system enhanced by artificial intelligence, which in real time detects the suspicious behavior of the client by 1000 different factors and gives a signal. After that, the employee checks all the identified matches with the "portrait of the drop," if necessary, conducts an additional investigation or immediately limits remote banking. This makes it difficult for fraudsters to receive and withdraw funds through transfers or ATMs.

Over the year, Tinkoff blocked more than 66,000 accounts that teenagers sold to attackers. An artificial intelligence-based fraud monitoring system helps to find them. The system analyzes the client's behavior and payment habits by a thousand factors. For example, it fixes where how much money a person transfers and where. This allows you to block cards at the first suspicious operation. For example, when tens of thousands of rubles began to be transferred from the teenager's account to suspicious details.

Drops have characteristic behavioral signs, the way they use the score and application or the fact that they transfer money to suspicious details. Thanks to the work of the system, it was possible to reduce the number of drops by 2 times in a year. And the proactive blocking of drop accounts by 2.5 times reduced losses due to their unfair activities, - explained Oleg Zamiralov, head of the Tinkoff Ecosystem Protection Center.

Thanks to the work of the system, we blocked more than 66,000 accounts that teenagers sold to attackers, "added Oleg Zamiralov.

To learn how to bypass bank protection systems, fraudsters create special groups for drops, in which they discuss ways to avoid blocking accounts. Tinkoff regularly identifies such groups and publics and sends them to block.