Badges of sellers of Tele2 salons listen to their talks with customers

Tele2 mobile operator the first on the Russian telecom market implements the intelligent audio system of the analysis of customer service in a distribution network, told TAdviser at the Russian office Microsoft on August 22, 2019.

It is necessary for system operation that sellers in salons carried audio-badges which help to recognize automatically their speech using artificial intelligence technologies. Their developer - Voca-Tech company, and as artificial intelligence technology is used the cognitive service Speech to Text Recognition from Microsoft placed in Microsoft Azure cloud. The Wi-Fi-module which is built in an audio-badge allows to transfer audiodata without additional devices. The accumulator ensures 12-hour functioning without recharge.

Such system allows to estimate the consultant's conversation with the client and to automate monitoring of quality of service.

Audio-badge of the consultant of Tele2 (a photo source - Microsoft)

The solution analyzes a speech transcription on business metrics, key for the company. Treat them: the solution of the main request, transition on MNP (Mobile Number Portability is transfer of the mobile number), informing on relevant promotion actions, acquaintance to a personal account, work with claims, etc.

Key technical problem which we solve – receiving qualitative audio recordings from a badge in any conditions, even at the high level of noise as, for example, in shopping centers or the airports, - Bella Mutallapova, the CEO of Voca-Tech notes.

At the initial stage of a metrics are configured by the person, but in process of self-training of neural network model processes are completely automated. Tele2 can supplement and replace business metrics depending on what characteristic in work of the seller requires improvement or on what offer it is necessary to focus promotion. A system allows to analyze quantitative parameters of the speech (how many times the seller said a phrase) and quality characteristics (what percent of visitors received the offer). It is possible to browse and manage analytical reports through a personal account.

In addition to business indicators, the intelligent system allows to estimate politeness, skill to communicate of the seller and existence of filler words. The solution helps to analyze quality of work of each consultant and to see the comparative ratings of employees. Tele2 is going to integrate the solution with the portal of training and to create the strategy of individual development of sellers.

It is possible to call a system an alternative to video monitoring and the "secret buyer" format which are often used in retail for job evaluation of selling assistants (photos are TASS/Vedomosti)

Where video monitoring cannot be organized, Tele2 looked at a task from a different angle, having entrusted speech analysis to artificial intelligence. The audio system with opportunities of speech analytics is thinly configured under business challenges and studies at a basis of neural network, – Denis Goleshchikhin, the director "says Tele2 Retail".

System implementation is planned within general strategy of digital-transformation of business of Tele2, and the decision on implementation was preceded by a pilot project. During it employees of the modules Tele2 in shopping centers received Voca-Tech audio-badges. Provide data into Microsoft that in retail outlets from the sale implemented by an innovation grew by 5%. By December, 2019 Tele2 is going to implement a system on 800 sales points across all Russia.

Microsoft told TAdviser that the Tele2 project is one of the first public business cases of use in Russia of cognitive service of speech recognition of Microsoft. The implemented solution on its base has clients only in Russia so far. At the same time as of August pilot projects in a number of the industries are conducted: finance, retail, builders, fuel companies, telecom, power, pharmaceutics.

In February, 2019 using the solution Voca-Tech in the departments announced VTB bank start of a pilot project. It was planned within the corporate accelerator organized by bank based on GenerationS accelerator from RVC.