The 2016 State of the Speech Technology Industry: Speech Analytics

Article Featured Image

Speech analytics had a very good year in 2015, and the reasons why weren't hard to assess. "In the past 12 months, vendors reached a number of milestones that resulted in these systems becoming much more functional," says Donna Fluss, president of DMG Consulting. That momentum is expected to carry over throughout 2016.

Historically, recognizing the spoken word has been challenging, but recently the systems have improved, so accuracy rates are rising. This can partly be attributed to vendors enhancing their algorithms so that more words are recognized. In addition, the development of speech-to-phrase solutions means that words are put into context, which increases understanding. For instance, the word angry has one meaning but if the word not precedes it, the meaning changes dramatically.

Such improvements figure into why speech analytics adoption has been so strong of late. The number of contact center speech analytics seats grew by 21.4 percent, from 2,889,031 in May 2014 to 3,507,795 as of March 2015, according to DMG Consulting. The market research firm expects the good times to continue with growth of 20 percent in 2015, 19 percent in 2016, 18 percent in 2017, 17 percent in 2018, and 16 percent in 2019.

A Large List of Suppliers

This rise is attracting a bevy of competition. Aspect Software, Avaya, CallMiner, Genesys, inContact, Interactive Intelligence, Nexidia (which NICE Systems recently agreed to acquire), Nuance Communications, OnviSource, Verint Systems, VPI, and Zoom International are some of the vendors selling speech analytics solutions.

Speech represents a good starting point for evaluating the customer experience because it is a very rich exchange. A five-minute phone call consists of about 5,000 words, much more than other channels, like chat or email.

But analyzing the spoken word has been expensive and time-consuming. In most cases, businesses would sample a few recordings and supplement that information with various surveys. "Businesses would sometimes gain inaccurate pictures, or false positives, from surveys and random samplings," says Daniel Ziv, vice president of voice-of-the-customer analytics at Verint. Improved speech analytics solutions now enable them to see all of the connections in every call.

Do You Feel Happy?

Emotion detection is another emerging capability. Here, speech analytics technology helps organizations determine a customer's mind-set. "What the company would like to know is what makes a call change from happy to sad or vice versa," Ziv says. Speech analytics accomplishes this goal by correlating words to actions.

Recognizing the impact of words is the first step in turning data into actionable information. Businesses then need to put processes in place to promote various responses. If the customer is in a good mood, then the agent can try to upsell the firm's goods and services. If the client is upset about the service, the business may offer a rebate or coupon.

Such insights can help other departments in addition to the contact center. Via Big Data and analytics, businesses mine customer interaction data and gain business insights.

Listening to and then analyzing a series of phone calls can unearth unseen problems. Say a computer manufacturer releases a new laptop that has software bugs. One call to the contact center will be glossed over, but a series of interactions will point the company toward the problem. The firm then needs to take that insight and integrate it into its operation. The business goes back to the product development team, outlines the problem, and crafts a solution. In addition, the executive team must determine how to assuage customer concerns. Customer service agents are apprised of the process and the resolution and pass the information along to customers. A perk, like a coupon, can be offered to customers to alleviate any ill will.

SpeechTek Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues