It determines when the borrower will repay the debt, and also the ways of interaction with it and their frequency. The system uses machine learning technology. It was developed in-house. The economic effect of the implementation according to the results of 2019 is projected to be more than 100 million rubles.

The program consists of two modules. The first predicts the probability of debt repayment by a client on the specific day of delay. For example, based on the individual data of the borrower it can calculate what part of the debt the client will repay for 20 days. “If we see that the borrower can pay the debt within 10 days, we minimize interaction with him, reducing the costs of messages and calls,” Irina Khoroshko. MoneyMan CEO explains.

The second module helps to choose the optimal channel of interaction with the client. The system determines the most effective way of communication with the client – via of calls, text messages, e-mail, etc. By selecting one channel, all others are disabled. In addition, the module determines the frequency of interaction with each client, reducing its volume. It is noteworthy, that all interactions with borrowers occur only remotely.

Both modules work in a bundle. The artificial intelligence system uses several hundred variables: user behavior and payment discipline. These are personal data, information about how often the user enters the member area, how he answers calls, the reasons for delay he/ she gives, in what installments he/ she repaid the debt, etc. If the client has already contacted the company before, the program analyzes his/ her behavior related to previous loans. Most often, the necessary data for forecasting are collected on 7-15 days of delay. The accuracy of forecasting is 87%. It will increase with each update of the program. The average debt collection level after the introduction of the program increased by 3 percentage points in 2019.

To create the program MoneyMan has conducted tests for several months. According to Irina Khoroshko, new technologies will make the industry more civilized: “The collectors in Russia have a bad reputation for a number of reasons: work of quasi-legal, unlicensed companies, short history of activity, low qualification of personnel, etc. We want to show that the debt collection could be not only ethical and conscientious, but also highly technological”.