VoiceAnalytics a 3-stage rocket

A 3-stage rocket to launch QM

Gartner’s estimate that 89% of companies compete primarily on Customer Experience (CX) highlights the importance of properly managing the quality of service delivered by contact centers to their customers.

The benefits of excellent Quality Management (QM) are apparent:

  • Improved customer satisfaction
  • Reduced customer churn
  • Improved agent satisfaction
  • Better trained and better-skilled agents

Quality Managers Challenges
On the other hand, the challenges for a Quality Manager are manifold. Excellent and efficient Quality Management starts with having a good and proper view on how contact center agents service their customers. That’s one of the biggest, if not the biggest challenge for a Quality Manager because it requires a very laborious and time-consuming process of playing back and listening to past conversations. Quality managers only will be able to do this for 2% to, at best, 3% of all conversations.

As a result, particular conclusions might be biased.

“My manager just picked a couple of conversations from the morning that my daughter was rushed to the hospital.”

Or particular issues might go unnoticed

“As a manager, I didn’t know that John, all too easily, told the customer to file a complaint.”

That’s where technology, in general, and VoiceAnalytics, in particular, comes to rescue.VoiceAnalytics acts as a 3-stage rocket for the Quality Management of your Contact Center:

Stage 1: It analyzes not just 2 or 3% of all conversations, but 100%
Stage 2: It easily automates the scoring of numerous quality criteria.
Stage 3: It learns from the feedback to further improve Quality Management.

Based on the analysis of 100% of the conversations, VoiceAnalytics can check specific quality criteria and use those results for Automatic Quality Rating.
Take a simple example of rating an agent’s correct and professional greeting of the customer.A couple of aspects jump to mind immediately when thinking of a professional way to e.g., answer an incoming service call.
The customer is greeted promptly when the call is picked up. Not after 5 or 10 seconds, because the agent was not paying attention or not ready to help the customer.
The agent introduces him/herself by stating his/her name and the company they’re working for
Finally, an agent asks politely what he/she can do to help the customer.

VoiceAnalytics configuration
VoiceAnalytics can very easily be configured to detect:
The agent starts his greeting with a proper “Good morning,” “Good afternoon” or “Good evening” and not just a “Hi” or “Hello.”

  • any silence at the beginning of a conversation and flag those conversations with silences that exceed a maximum allowed duration (of, e.g., 5 seconds)
  • the expressions Good morningGood afternoon or Good evening and flag those calls where none of these expressions were used at the start of a greeting.
  • agent and company names. VoiceAnalytics supports the addition of custom-defined keywords and, as such, supports keyword spotting for out-of-dictionary words.
  • particular sentences such as How can I help you today? VoiceAnalytics supports the definition of key parts of a sentence: howhelp, and today, and defining multiple variations of a particular sentence, such as What can I do for you today?

VoiceAnalytics then uses its detection of the presence or absence of the above aspects to rate each conversation for this specific quality criterion.
A call in which the customer was greeted promptly with Good morning, you’re talking to John Doe of Company X. How can I help you today? Will automatically get a score of 100% for the Proper Greeting quality rating.

While a call in which the customer is greeted promptly with Good morning, I’m John. What can I do to help you today? Will automatically get a score of 75%.

And a call in which the customer was greeted after 10 seconds of silence with Good evening, uh… this is John will automatically get a score of 25%.

Customer Experience and Quality Management
By implementing this Automatic Quality Control, time is now freed up for Quality Managers to tackle the most challenging and subtle aspects of Customer Experience and Quality Management.
Take the example of managing customer frustration. Quality Managers can play back and listen to conversations and judge whether an agent can properly manage a customer’s frustration and turn the customer into a satisfied customer at the end of the call. While it’s extremely difficult for a Quality Manager to define frustration management in terms of words, expressions, and speech characteristics that should have been used by an agent, the last stage of the VoiceAnalytics rocket can determine that for them.

Based on the Quality Managers’ classification of calls with and without proper frustration management, VoiceAnalytics’ Machine Learning capabilities can determine typical call characteristics like speech characteristics, words, expressions, sentences, or even CRM data that correlates to what was classified as good frustration management skills.
As a result of that, rules can be defined to allow the Automated Quality Management system to classify the agent’s behavior as good or bad for managing a customer’s frustration.
As Quality Managers classify more calls, more input is given to the Machine Learning engine that can then further fine-tune these rules.
This approach can even be extended to create a self-learning system.
By merely providing, e.g., the NPS together with each recording for which an NPS was given, VoiceAnalytics’ Machine Learning capabilities can determine the rules for automatic scoring of those calls for which no NPS was given. The continuous stream of new call recordings with their associated NPS can then be analyzed extensively, and the results of that pattern and correlation detection analysis can serve as input to further update and fine-tune the NPS rules over time.

As a result, the system becomes self-learning.
A very similar example can be worked out for creating rules, and self-learning cycles for agents successful in upsell opportunities by providing the VoiceAnalytics’ Machine Learning engines with an indication of a successful or unsuccessful (up)sell attempt for every conversation.

VoiceAnalytics acts as a 3-stage rocket where the first stage of the rocket will boost the percentage of inspected conversations from 2 or 3% to 100%. The second stage of the rocket will transform Quality Management from, for the most part being a manual effort to, mainly being an automated process.
Finally, the third stage of the rocket makes the analytics solution learn from feedback to create a self-learning system ultimately.

Ignite your VoiceAnalytics rocket NOW … 3, 2, 1 … and take-off … to the next level of customer experience!

Latest News