Use of speech analytics: What is the best business model?

The use of speech analytics in call centers is increasing but many people are still wondering whether to use it or not. For those who do not know it yet, speech analytics is nothing more than an automated way of analyzing the content of the calls.

use of speech analytics

In the case of voice calls, the recordings must be transcribed by a transcription mechanism prior to be converted into text and thus, a computer program, automatic and intelligent enough to search what we want to find out without intervention human, can analyze them.

Different to what many people think, the Speech Analytics is not only used to monitor more calls and get a more accurate quality index. This is not the real benefit of this tool.

With the use of Speech Analytics, organizations can analyze handled calls by finding offending words, synonyms, phrases, sequences, business opportunities, mentions of competitors and brands, specific needs of customers etc.

It can also be analyzed adherence to the scripts, presence or absence of important procedures, such as confirmation of address, social security number etc. It is all a matter of creativity and knowing of how to use the tool, which with its own use, day by day becomes “smarter”.

Compare the use of speech analytics and its business models:

Monitored by Methods and Goals
(Quality monitoring team)
  • The quantity of monitored calls (usually does not exceed 1% of the total handled calls)
  • Finds errors in processes
  • Finds errors in information bases
  • Evaluates skills of agents
  • Quality assessment (approximate, based on the average quality index of the sampling)
  • Furthermore comprehensive analysis
Speech Analytics

(computerized process,
without human intervention)

  • Performed by sampling, but it can reach up to 100% of the handled calls
  • Finds words or phrases that are offensive or may cause loss
  • Classifies handled calls
  • Automates the monitoring by automatically filling in the checklists (monitoring forms)
  • Evaluates adherence to scripts
  • Analyzes the more efficient arguments
  • Security: used for fraud prevention and other security rules (applied to 100% of the handled calls)

(process carried out by the quality monitoring team with the support of speech analytics)

  • Assists the monitors by partially filling out the checklists
  • Selects which records to be sent to the quality monitoring team (or auditing team)
  • Filters and separates records for further analysis of sales techniques, achieving results, business opportunities, etc.


It should be noted that computers still cannot replace the analytical capacity of a qualified professional with huge product and business rules knowledge, but it would not be feasible to manually analyze a large volume of handled calls. For this reason, if your need is to analyze a large volume of handled calls (or even 100% of them), the shortcut seems to be Speech Analytics.


The implementation of the hybrid business model is a great solution for organizations that have a large monitoring team and at the same time need to increase the quantity of monitored calls. In this way, more efficiency can be achieved with an equivalent or even less running costs.

Know the CYF Analytics Solution.

The CYF Analytics work with IBM Watson and others providers.

Leave a Reply

Your email address will not be published. Required fields are marked *