The Best Way for Contact Centers to Deploy Speech Analytics
November 10, 2017
By ASC Horizons
Since all but the smallest contact centers deploy speech analytics, their business success will depend upon how efficiently they use this remarkable solution. The ability to analyze every customer call has become powerful and indispensable for many industries, and it has transformed the contact center into a valuable source of real-time business intelligence on an enterprise-wide basis.
Organizations proficient in IT development will handle the adoption of speech analytics like any other new technology. They will create clear goals, establish qualified individuals and processes, and then measure, measure, measure! And the deployment of speech analytics should follow a similar pattern.
Conducting a Demand Analysis
The first step in creating a clear goal involves conducting a demand analysis of customer interactions. The contact center must know why customers are calling – and what is normal – before monitoring interactions to discern burgeoning trends, analyze marketing campaigns or discover troubling conversations.
Keyword and phrase spotting enables a content-based call selection. A successful demand analysis will use this content to categorize 90-to-95 percent of interactions. If the percentage of uncategorized calls rises above 10 percent, then the keywords themselves should be adjusted accordingly.
The most complex form of speech analytics, transcription-based analytics, creates a text version of the phone call. While it enables data mining and the search for words without pre-selection, it requires voluminous dictionaries, continuous adjustments and a tremendous amount of computing power.
One of the most fundamental uses of a demand analysis involves simply counting the number of calls in each category. The ability to detect spikes in specific types of calls, that is, calls containing unique keywords, can provide near real-time, unbiased feedback on the interface between customers and the company purporting to serve them.
If the number of uncategorized calls increases significantly, it may indicate a new trend or a subject currently unaddressed by the enterprise. Operational changes such as website modifications may be causing confusion. Or maybe the company’s product contains a flaw or is missing a feature the customers demand.
The Test-Improve-Test Process
The best use of speech analytics on a daily basis involves implementation of the test-improve-test process. When the demand analysis and simple call counting indicate a problem, management must determine the best way to ameliorate the situation.
Then, whatever fix is implemented can be evaluated by determining its impact on the call categories. If nothing changes, or the improvement needs to be supplemented, then speech analytics can verify the impact or suggest other modifications. Management gets scientific, quantifiable feedback on the effectiveness of their decisions.
Techniques such as demand analysis, counting calls and the test-improve-test process represent powerful techniques for the deployment of speech analytics in a contact center. They will help to optimize your speech analytics solution and ensure satisfaction among your customer base.
About the Author: Head of PreSales for ASC (News - Alert) since March 2015, Uwe Kleinschmidt develops global business solutions, manages and structures the PreSales team, and adapts their activities to the latest market conditions. With more than 20 years of experience in Unified Communications (News - Alert) and contact centers, he has focused on sales support, consulting, system integration and solutions design.
ASC Technologies AG is a leading global provider of innovative solutions to record, analyze and evaluate multimedia-based communications. ASC’s solutions reveal vital information, enabling companies and organizations to considerably improve their value creation: contact centers enhance customer service, efficiently deploy staff and increase productivity. For more information, visit www.asc.de; call +49 6021 5001 0; email [email protected]; or write ASC Technologies AG, Seibelstrasse 2-4, 63768 Hösbach, Germany.
Edited by Erik Linask