Sber Robot trainer for training bank employees in communication with debtors

Main articles:

• RPA (Robotic process automation)
• Overdue loans in Russia
• Speech technology: from recognition to understanding

2021: Creation of a robot trainer in the field of collection of overdue debts

On September 30, 2021, Sber announced that he had created a robot trainer for his employees working in the field of collecting overdue debts. Such an algorithm, acting as a client, is a full-fledged combat simulator that helps an employee improve communication skills with debtors.

As the director of the division "Retail collection and settlement" of Sberbank Denis Kuznetsov noted, the use of robotic communications in the field of collection is growing every year. The first modification of the "robot collector" was launched in 2017, and already in 2018, technology based on artificial intelligence conducted 7% of dialogues with customers. In 2020, the share of robotic dialogues increased to 77%, and by the end of 2021 it is projected to exceed 85%.

The capabilities of the algorithm, which are sometimes indistinguishable from human ones, are constantly growing. So, the robot has already learned to adjust to the speed and volume of the client's speech, if the client speaks quickly or asks the robot to speak faster. The robot knows how to use the history of communications and can remind the client that he promised to pay the debt and has not yet done so. The algorithm is able to warn customers about the upcoming payment and answer clarifying questions about the amount and date. And finally, in October 2021, the robot will begin to send SMS at the request of the client at the time of the dialogue. This will allow customers who request additional information to receive it directly during the conversation. All of the above robot knows how to do in relation to representatives of small businesses.

Шаблон:Quote 'author = said Denis Kuznetsov, director of the division "Retail collection and settlement" of Sberbank.

As Denis Kuznetsov emphasized, research shows that customers are calm about communicating with the robot, understanding why this is done. In addition, this format is used everywhere in various service industries. Speaking with the robot, people do not experience feelings of guilt or inconvenience, which, as a rule, arise when communicating with a person when discussing debt. Promises to pay off the debt that the client makes to the robot are kept with the same probability as promises made to the employee.

Since 2018, the bank has been conducting regular measurements of the customer satisfaction index with telephone negotiations on debt collection, and in the last two years the robot has consistently received higher ratings than the operator. Arguing their assessment, customers usually talk about the weighting of the robot, sequence and clear argument. Thus, the prospects for further robotization of the industry as artificial intelligence technologies develop look positive.