Kevin Hegebarth, Director of Strategic Planning, Witness Systems

Q.  Why is speech analytics important to enterprises?
A.  Every customer contact potentially contains a wealth of information.  During these interactions, customers divulge information about their satisfaction or dissatisfaction with a company's goods and services, provide valuable insights into a company's existing and emerging competitors, offer feedback on a company's marketing campaign and image and offer clues into whether their interactions with the organization's call center agents did or did not satisfactorily address their issues.

With the sheer volume of calls handled by many call centers today, it's simply not possible for managers to listen to each and every conversation to glean these insights.  Further, the data provided by a call center's CRM, ACD, CTI and other systems provides only part of the "big picture."  For this reason, speech analytics technology serves an important role, enabling enterprises to gauge customer satisfaction, competitive pressures, and compliance to policy and procedures.  Having speech analysis software effectively listen to all calls recorded by the enterprise, classify those interactions based on what was said in the conversation, and provide trend data over time is like having thousands of supervisors working around the clock.

Sometimes enterprises know what types of calls they wish to analyze, but they may not know what speech and language patterns best satisfy the analysis criteria.  It's therefore essential that speech analysis also include the capability to mine the recorded conversations to uncover what key words and phrases were most frequently used, even if those words and phrases are not known ahead of time.  This provides an "early warning system" of what's being said that an organization may not have knowledge of in advance.

Speech analysis provides an "outside-in" perspective on a company that may not otherwise be apparent by conducting samples of agent evaluations or customer satisfaction surveys.  It also offers an important complement to agent evaluations and customer satisfaction surveys by collecting information about interactions as they occurred and rationalizing them against those customer surveys and agent evaluations.

Q.  Where do you see this technology going over the next two years?
A.  Natural language speech recognition technology continues to improve in terms of recognition rates and processing speed, which will result in higher confidence results and the ability to process a greater number of interactions without massive hardware expenditures.  This will serve to improve the speech analytics applications built using this technology.

In the near term, three applications appear to have the greatest promise: 

  1. Query-response for pre-determined words and phrases.  Sometimes referred to as word spotting, this application is simply a user, either through a suitable user interface or through some business rules engine, querying the database of recorded interactions for a set of known words and phrases.  The result could be presented as a report or collection of interactions that best meet the search criteria. 
  2. Proactive analysis.  Uses the power of the speech engine to return to the user the most commonly used words and phrases found by analyzing the recordings.  This is helpful in uncovering previously unknown speech trends and can be used to quickly thwart a competitive threat, re-tool a marketing campaign or launch a customer retention program.  Such proactive analysis also can act as a catalyst for creating new product offerings, price bundles or marketing campaigns. 
  3. Trend analysis.  Identifying the relative frequency of certain words or phrases can give enterprises, for example, valuable insight into what their competitors are doing.  For instance, a change in the relative frequency with which a specific customer's name is mentioned, or the relative frequency with which price is mentioned as a reason for terminating service, can alert enterprises of the need to take action quickly.  

Q.  Are there particular vertical markets you believe will adopt speech analytics sooner than other verticals? Why?
A.  Speech analysis has broad applicability across all verticals.  Industries that perform the bulk of their customer interactions over the phone and are highly competitive - such as airlines, wireline and wireless communications, and financial services - are good early adopter candidates for this technology.

Q.  How can this technology improve a company's marketing?
A.  The effectiveness of marketing campaigns is often measured by increased sales volume.  Many times their success isn't known until the campaign is over and all the numbers are in.  Speech analysis can provide an up-to-the-moment view of what customers think, how well a campaign is being accepted, and what, if any, adjustments should be made.  Speech analysis provides insight into product, pricing, packaging and promotion outside of targeted campaigns.  In effect, speech analysis can become relatively inexpensive market research.

Q.  Tell us about Continental Airlines experience with your eQuality CallMiner product.
A.  Continental Airlines, which employs more than 900 customer service representatives in its Tampa center, classifies customer calls into more than 50 categories for analysis, including sales calls, flight information, seat assignments, reconfirmations and OnePass reward redemption.  Using Witness Systems' speech analytics solution, it's able to reduce the costs associated with compiling its "Call Mix Survey" - a system that collects and tabulates detailed information on incoming calls at the company's reservation center - by automating the process, saving significant time and resources. 

The speech technology software will enable Continental to more easily categorize customer interactions by call type.  More importantly, it will provide metrics about which customers call the reservation center and why.  This call classification process also will assist with decisions about which calls to direct to its Internet, self-service or automated speech system channel, reducing expenses in the reservation center and freeing call center agents to handle high-touch customer service and high-value, revenue-generating calls.  Further, management will be able to quickly review a much larger sampling of interactions, thereby gaining the key metrics and broader perspective necessary to drive strategic business objectives. 

Q.  How long did it take to implement this technology at Continental?
A.  The implementation of the technology is in progress presently.

Q.  What have they learned from implementing CallMiner?
A.  Continental has been a Witness Systems customer for a number of years, leveraging the company's customer interaction recording and performance assessment software solutions and experiencing impressive results - ranging from increases in e-ticket sales and vendor transfer programs, as well as increases in first call resolution, productivity and quality scores.  By extending its technology deployment and adding speech analytics to the mix, the airline looks forward to gaining more context and intelligence from a speech analytics perspective, enabling it to better leverage the customer data and feedback it already captures in order to make more informed decisions.  In the past, Continental completed monthly Call Mix Surveys, but now it will be able to review the data daily, which will provide more real-time insight into what's happening in the business and better enable the airline to focus on service excellence.

Q.  What is different in your approach to speech analytics versus your competitors?
A.  Many of our competitors emphasize word spotting as their speech analysis solution.  With this approach, the user has to know ahead of time what key words/phrases to be looking for.  Many valuable opportunities can be lost using this "looking in the wrong place" method.  Implementing all three applications:  Query-Response, Proactive Analysis and Trend Analysis provides the enterprise with the best opportunity to assure customer satisfaction; quickly and effectively respond to competitive threats; ensure call center agents are trained effectively; and facilitate an outside-looking-in culture.

Q.  Any last thoughts?
A.  Speech analysis is an important component in any company's workforce optimization, customer satisfaction and retention, and revenue generation strategy.

SpeechTek Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues