VTB will feed artificial intelligence the entire "voice of the client" in order to improve the quality of service

As TAdviser discovered, in its sustainability report released at the end of June 2025, VTB Bank announced a new large-scale project - "Voice of Client 2.0." It was initiated as part of the customer experience quality strategy. Quality management of customer experience is one of the key areas of the bank's business strategy, designed for 2024-2026.

Within the framework of the "Voice of the Client 2.0" project, data on the "voice of the client" will be consolidated for the first time - client contacts with the bank in all channels: calls to the call center, visits to branches, chats with the operator and bot, surveys on the quality of client experience, complaints and complaints, etc.

The initiative will involve more than 30 streams and divisions of VTB Bank and 80 IT development teams.

With the help of its own machine learning models and elements of artificial intelligence, the client's voice will be analyzed taking into account all knowledge about the client, the history of its service - in order to constantly provide divisions and streams of the Retail Business unit with online information about customer needs.

𝓲

Фото: Fullvector/Freepik

In addition, within the framework of the project, new client ways of collecting feedback will be implemented on the basis of VTB Online and the website, as well as the centralized "employee voice" about operations performed for customers in branches, call center and remote sales services, including delivery.

Useful analytics with client voice data from various channels will be deployed on a modern analytical BI platform that takes into account import substitution criteria.

The first results of the project, including the creation of the first ML models for "reading" complaints and complaints in a machine way, were achieved already in 2024, VTB Bank reports in its sustainability report. And during 2025 it is planned to fulfill the bulk of the tasks.

As for improving the collection of "client voice," 2024 year, for example, 120 new functions were introduced into the chat and chat bot at VTB Online, artificial intelligence appeared in the chat bot, and the number of personalized scripts increased 2 times.

Due to the development of the chatbot, the share of chat from all remote calls per year increased from 44% to 61% - more than half of customers choose chat to resolve issues. And the introduction of machine learning has improved request recognition, increasing automation to 81% and the share of satisfaction to 80%, according to VTB Bank.

To understand the scale: the number of VTB Online users in the reporting period reached 22 million people, and the VTB contact center receives more than 200 thousand customer calls every day. At the same time, in one appeal to the contact center, the client asks from one to five questions. For efficient work and high-quality service, all requests are classified. This allows you to route the client to the desired operator from robotic channels (IVR, chatbot), automatically launch a service script when the client communicates with the operator, carry out online monitoring and load control of the contact center.

Autoclassification also helps to analyze the dynamics of calls: channels, topics, regions, periods, etc., in order to give feedback to back-end departments to improve products and services. As of mid-2025, more than a thousand topics in the service voice channel and more than 700 topics in the chat have been classified. Thanks to the online classification in 85% of calls, where the client voiced his question, the topic was determined successfully, and the desired scenario was automatically launched in the operator's workplace.