Speech Analytics Gets Real (Time)

Article Featured Image

Industry Data Security Standards (PCI-DSS), which is intended to ensure security when agents are processing credit card information. Collection agencies also have to advise or warn debtors—what's known as a mini-Miranda—that they are attempting to collect money and that the call could be recorded. In both the payment and collections scenarios, speech analytics can be used for quality assurance (QA) purposes to ensure that contact center agents are adhering to compliance rules.

Although speech analytics is primarily used for productivity in contact centers, companies are beginning to apply insights from recorded calls to other departments too. For instance, a customer on the phone with an agent might state a competitor's name or product or raise his voice in frustration, which can also be discovered using a decibel-level search. Speech analytics could capture that information, and, in turn, those insights can become actionable if used by other departments, such as marketing.

"Keep in mind that speech analytics is capturing everything that customers say," Fluss states. "Whether they're talking about product ideas or having trouble understanding instructions, speech analytics should be shared and used in all parts of an organization."

Post-Call Speech Analytics: Buried Gold

The difference between post-call and real-time speech analytics boils down to the point in time in which speech from a call is captured and processed in the contact center. Post-call speech analytics listens and analyzes calls after they have concluded. Real-time speech analytics occurs simultaneously as a caller is speaking to a customer service agent. Both technologies have many advantages.

With post-call analytics, a recorded call is fed into a processing engine and then indexed, D. Daniel Ziv, vice president of voice of the customer analytics at Verint, explains. Additionally, post-call speech analytics solutions may apply natural language, which Ziv says provides better understanding of context. The resulting data can provide automated insights and allow users to search against an index for insights. The data can be sliced and diced by different factors and offer a look into areas such as agent performance or dissatisfied customers.

"Post-time speech analytics solutions can automatically identify root causes, trends, and emerging trends," Ziv says. "It's been around for a while and is much more mature in the sense that we're on the fifth or sixth generation of this. Verint's first commercial deployment was in 2003, and there [have been] many hundreds and hundreds of successful deployments since then."

Additionally, for post-call speech analytics, complete accuracy is not as important as it is in real-time speech analytics, since the software is mining a large number of calls and then transcribing and indexing 100 percent of all identifiable words and phrases.

"Missing several individual words typically makes little impact when you're trying to identify statistically significant trends dealing 

SpeechTek Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues
Related Articles

Nexidia Brings Neural Phonetic Speech Analytics to Audio Discovery

Enhancements provide visual data analysis and improved tools for review and workflow for the legal market.

Bedard Law Group Deploys Nexidia's AudioFinder

Nexidia AudioFinder will help the law firm and collections agency ensure compliance.

Analytics Revolutionizes the QA Process

Increased targeting and precision make an upgrade worthwhile.

Noble Systems Adds Real-Time Screening to Speech Analytics

Real-Time Speech Analytics can provide agents and supervisors immediate feedback during the call.

Star Performers: Nexidia's Analytics Addresses 'Enormous Volume and Scale'

Companies and Suppliers Mentioned