Skip to content

How is analytics-enabled QA different than traditional QA?

11/5/2014

  Printer Friendly Format    

Question
How is analytics-enabled QA different than traditional QA?

Answer

All contact centers, regardless of size, number of sites, number of agents or channels used, should perform quality assurance. At its most basic, quality assurance is a process designed to measure how well contact center agents adhere to internal policies and procedures. When done properly and on a timely basis, it also functions as an early warning system to identify trends, issues and opportunities for all areas of the company.

  
Today, there are two primary methods of conducting quality assurance, traditional and analytics-enabled. Traditional QA is a labor-intensive function. Supervisors and quality reviewers search through hours and hours of recordings to select a subset of recordings for evaluation. In some cases, QA applications can qualify calls or email/chat interactions for quality evaluation based on pre-defined criteria such as call direction, call duration, interaction type (based on wrap-up, disposition or other interaction classification mechanism), product type, program, etc. Others may use business rules to identify interactions that require attention. However in both of these cases, while the QA process is facilitated and supported by technology, it is never truly automated. And, because much of the process is manual, only a small, statistically insignificant percentage of interactions are reviewed and evaluated. (Many contact centers receive millions of calls each week, but only have the staff to evaluate 1 to 10 interactions per agent or 1% to 3% of all interactions, if they are lucky.) Although traditional QA is not ideal, for the past 30 years, it was the only way to provide some form of oversight, let agents know that they were being monitored, identify a small portion of performance issues, and surface operational, procedural and system trends for the company.
  
Analytics-enabled QA is the antithesis of traditional QA. Analytics-enabled QA solutions leverage the capabilities of speech, text and desktop analytics to automate and improve the QA process by making it more targeted and precise. First, analytics-enabled QA solutions use speech analytics to review (and in some cases, score) as much as 100% of calls in order to identify interactions that require attention. Speech analytics can be used to identify calls that require attention based on other KPIs such as being out-of-compliance with an established script, calls that include high emotion, talk-over and/or positive or negative sentiment, or calls that contain/do not contain specific key words, phrases or concepts. Speech and text analytics can also be used to identify broader enterprise trends, such as whether customer attrition is an issue, if a competitor has a new deal, or if there is a system or process that is causing high levels of customer dissatisfaction. Desktop analytics adds another dimension to the QA process; by providing visibility into what agents do at their desktops, including the fulfillment process, it can “watch” to see if agents are delivering on their promises to customers, such as by accurately processing monetary and non-monetary transactions.