In Russia, an AI Assistant Is Set to Audit Contact Centers and Grade Their Performance

Yandex B2B Tech has opened access for businesses to AI assistants capable of automatically analyzing conversations between operators and customers. The neural network evaluates service quality and identifies weak points in operator performance.

Analyzing Thousands of Conversations

The new tool is built on the Yandex SpeechSense speech-analytics service and is designed for analyzing contact-center calls. According to the company, more than 30 organizations launched around 700 AI assistants during a two-month testing period. In large contact centers, this can save hundreds of work hours that would otherwise be required to process thousands of conversations each day.

The AI assistants analyze operator–customer conversations and draw conclusions based on accumulated data. The system automatically assesses customer satisfaction levels, generates concise conversation summaries, and then offers recommendations on how to improve service quality.

Different Assistants for Different Tasks

Businesses can create multiple AI assistants for different operational scenarios or use ready-made templates. Available options include quality control for customer support, analysis of active sales calls, evaluation of offline conversations at retail locations, and analysis of HR communications. This approach allows companies to tailor analytics flexibly to specific internal processes.

Elena Belobrova, head of ML services business development at Yandex B2B Tech, said the goal of the service is to make communication analysis as simple and fast as possible.

According to her, product managers can quickly identify which app features generate the most questions, while sales leaders gain tools to assess the effectiveness of promotions and marketing offers.

Proven in Practice

The new AI assistant is already being used by insurance and service companies, including major players such as Renaissance Life Insurance and the Urent (YouRent) kicksharing service, which is part of the MTS ecosystem.

At Renaissance Life, the neural network automatically determines customer satisfaction levels after contact-center interactions and compares the results with actual feedback. This eliminates the need to conduct customer surveys that often frustrate users.

Urent uses AI to analyze customer sentiment at the beginning and end of a conversation, determine whether an issue was resolved, and calculate the likelihood of repeat inquiries. The system also classifies requests by topic, helping identify the most common user problems.