Tochka Bank has learned to read clients' emotions using AI

2023: Customer Emotion Detection Service Launched

Point Bank on June 1, 2023 announced the launch of a service based on artificial intelligence, which determines the emotions of customers in a chat with the bank, monitors their dynamics and understands what is associated with the change in the mood of entrepreneurs.

The ML model, which was trained by Tochka, analyzes the client's mood in three main categories: negative, positive and neutral. Artificial intelligence assesses the client's mood throughout the conversation, drawing on context and emotionally-colored words. This helps the bank's employees not only to assess the general mood of entrepreneurs, but also rather to pay attention to those points that worry customers right now.

{{quote "The number of clients of the Point is growing every day, as of June 2023, we serve about 550 thousand entrepreneurs. It is important for us that all customers continue to receive the best service and be heard. The ML model helps us to find a special approach to each client, while not loading him with additional questions. So we find out the preferences of the entrepreneur in working with the bank, saving his time. Soon, the ML model will recognize the voice of the client. This will help us to catch even a "slight chill" emanating from a person in order to find an individual approach to him, - explained Anton Shilin, leader of the "Communication with a Client" direction at the Point. }}

Machine learning also helps to draw attention to the quality of technical support in the context of each expert and to increase the emotional value of each dialogue. Artificial intelligence and its own model of emotions will allow the company to respond faster to feedback from customers, focus on their needs and develop the bank, taking into account the opinion of each entrepreneur.

Recently, the number of requests from customers of the Point has quadrupled. Using the ML model, specialists track any change in client emotions. As a result, the number of dialogues that began with anxiety and ended on a positive note grew by 20% compared to the beginning of the year.