Artificial intelligence makes it easier for banks to deal with debtors, as the machine learning algorithms have enabled them to make more accurate projections about the client’s behavior and distinguish “bad” borrowers from “good” ones.

Representatives of Russian banks have shared with Izvestia their experience in applying AI for contacts with borrowers.

********

Home Credit Bank uses behavioral data for decision making. It used to be just several dozen indicators, but now their amount reaches hundreds and even thousands. Machine learning allowed the bank to make more accurate projections about loan default, consequently pushing the level of past due debt down to 3.7%.



********

B&N Bank applied AI to decide if past due payments in early stages warranted calling the client. The technology revealed 70% of cases did not require a call, which subsequently increased loyalty of the clients and did not affect efficiency of debt collection. Moreover, the bank saved money on penalties. With the help of artificial intelligence the bank identifies the clients who don’t want notifications and those who need reminder and consultation.



********

Tinkoff Bank uses AI for speech recognition and issuing of mortgage loans. The advisory engine makes 80% of all decisions on mortgage deals, which not only excludes the need of managers and saves money on operation processes, but also enables the bank to scale resources.

The partner of Fintech section is