Results
36%
retail sales growth
2.4 times
reduction of time of the head on control of the employee
2 TIMES
improving the accuracy of pronouncing company standards
Recognition accuracy
How the platform works
EMPLOYEE SPEECH ANALYTICS

We record and recognize speech, and then provide daily targeted feedback to the employee based on the words they say.

 

The platform sends e-mail and sms at the end of each working day.

The employee independently forms key skills within 2-3 weeks, instead of the usual 6 months.

This is how he principle of accelerated self-learning is formed.

CUSTOMER SPEECH ANALYTICS

All recorded dialogues can be automatically divided into themes.

This means that you can only listen to those that contain the information you need.

For example, you can find out:

  • customer claims on quality
  • suggestions for delivery
  • sources of advertising appeals

etc.

This analysis dramatically reduces the cost of the contact center and allows you to quickly make management decisions.

MEDIA ABOUT US
OUR PARTNER
USEFUL INFORMATION

We divide the implementation into three main stages:

  1. Acquaintance with technology

    In order to understand how speech recognition works, just send us recordings of telephone conversations or request free testing of audio badges.

  2. Implementation

    Using the service begins with a pilot implementation for a period of 2-3 months. In the process, word sets are formed, data transfer to the platform is debugged, and skills for working with reports are formed.

    Advantages of using the service:

    • Objective assessment of employees for compliance with corporate culture
    • Measuring the accuracy of the pronunciation of the script in %
    • Assessment of the effectiveness of staff training
  3. Service Implementation

    The implementation period of the service is from 3 months, it includes:

    • Integration and debugging of data transfer
    • Formation of word sets
    • Calculation of performance indicators of service use
    • Improving employee performance

    The average term for fully customizing and adapting the service for a particular business is 11 months.

Word sets help to adapt the platform to the specifics of the business. 3 stages of forming word sets:

 

Primary

Formation of word sets in accordance with current scripts.

 

Secondary

Employees supplement word sets according to speech formulations when communicating with clients.

 

Based on N-grams

N-grams are stable phrases that are pronounced by company employees.

The service divides the recognized speech into words and phrases.

A report on N-grams is a set of phrases of two, three, four and five words, which are sorted by frequency of use.

 

The report allows you to combine phrases into sets of words. The word sets are complemented by phrases that are used by employees, and those that are worth adding to use in speech.

Fonemica monitors the needs of the market, so it has developed a stereo voice recorder badge that helps analyze both the speech of employees and customers. Unlike a few analogs, 2 audio tracks from 2 microphones are recorded. As a result, we get a separate speech between the employee and the client.

When a badge is placed on the charger after changing the employee, the records are transferred to the analytics platform. For information security, records are encrypted and recognized only by the program.

In the body of the stereo voice recorder badge there is a light indicator that signals the conduct of audio recording.

The badge is charged wirelessly through the dock.

When processing telephone conversations, recognition accuracy depends on:

  • recording format: WAV, mp3, mono, stereo
  • compression settings
  • headset quality

When processing conversations with a badge, recognition accuracy depends on:

  • external noise: the presence of production, air conditioning, cars
  • polyphony: the presence of more than one voice in the room
  • employee speech: how intelligible, loud and clear

The audio recording is carried out by a full-time employee, but only those segments where there is speech are recognized.

To do this, the service removes silence and recognizes only "useful traffic".

It is difficult to determine how much retail employees communicate with customers, so the amount of recognized traffic is determined by the customer.

For full conclusions about the work of employees, at least 30% of conversations are analyzed: selectively, in a row or only during high-pressure hours.

According to statistics, in contact centers the quality control group listens to less than 2-3% of conversations, and in retail the efforts of the mystery shopper reduce this figure to zero.

Automatic transfer of conversations to the platform is configured by our technical support.

To request integration, send an email to info@fonemica.ru

Analytics is possible for all types of Stereo Voice Recorder Badges.

Separation of records by type and by employee occurs on the customer side: incoming, outgoing, by id.

Records of conversations in stereo mode can be analyzed separately for the employee and client. The stereo format allows you to build statistics on customers: how did you find out about the company, what needs are most often addressed, etc.

need advice?
fill in the application and we will call you back
By clicking on the button, you consent to the processing of your personal data on the terms and for the purposes defined in Privacy policy.