New Support-Center Tool Detects Emotion In Voice Of Disgruntled Callers

Software automatically alerts supervisors when customers voice frustration about company's goods and services.

InformationWeek Staff, Contributor

March 25, 2005

4 Min Read

Not long ago, an elderly man distressed over high medical premiums phoned the call center at Wisconsin Physician Services Insurance Corp. The caller was highly emotional, making it tough for the call-center agent to calm him so his problem could be addressed. The frustrated policyholder hung up. But the call-center IT system was aware of the customer's exasperation, and automatically fired off an E-mail alert to the agent's supervisor, who immediately listened to a digital recording of the conversation.

Moments later, the supervisor called the customer, and suggested ways to lower the premium--by increasing the deductible, for instance. The customer agrees to the policy changes. "All in all, we ended up with a happy customer," says Sharon Whitwam, the insurer's VP of member services.

Keeping customers happy is crucial for most businesses, and knowing when they're disgruntled is important to the Madison, Wis., health insurer. Last year, WPS began using new software that provides this insight. The software is called Perform and was created by call-center software provider Nice Systems Ltd., an Israeli company, which began widely marketing the product this month.

How does emotion detection work? Using algorithms, which Nice Systems spent nearly $30 million to develop, the system determines a baseline of emotion during the first five to 10 seconds of a call, when most people usually aren't excited or frustrated. Any deviation from that baseline, based on 26 parameters, can trigger an alert.

The software conducts a flow analysis of each call, examining 200 elements to give a holistic picture of the customer experience. Besides emotion detection, Perform uses two other methods to analyze customer experiences with call centers. The software allows users to create lexicons of words and phrases a caller may say that could raise red flags: cancellation, frustration, or a competitor's name. Perform also tracks the history of a call: the length of time a customer had to wait to speak to an agent, the number of times placed on hold, and the number of times a caller is transferred. The user sets up parameters to determine who and how a supervisor should be contacted. The system also allows callers to rate their experience with the agent after the conversation ends. Nice integrates with Microsoft .Net client-server technology, as well as most enterprise-resource-planning and customer-relationship-management systems.

The voice-quality systems monitoring market reached about $800 million last year, says Eyal Danon, Nice Systems' senior VP of sales and marketing, and he expects it to increase by 15% to 20% this year. In following years, he predicts, triple-digit percentage growth should occur as more businesses learn about the technology. He says 65% of Fortune 500 companies use Nice Systems technology.

A major attraction of emotion-detection technology is its promised ability to help identify customer frustration, a bane to executives who fear customers might take their business elsewhere because of poor customer service. A recent study by the Customer Care Alliance and Arizona State University suggests that nearly three-quarters of Americans have experienced customer rage when talking to a call center, with half swearing they'll never do business with that company again.

Voice analytical systems also are being exploited to learn more about customers and employees. Wisconsin law annually requires insurers to notify by mail Medicare supplemental policyholders about their policies. These letters contain technical terms and legalese, prompting calls from some muddled customers about their meaning. WPS uses Perform's word-spotting feature that identifies common words callers use, such as "I'm confused," "I don't understand," and "Why am I getting this?" Reports generated by these calls are forwarded to marketing experts, who also listen to some of the chats. This helps them create better ways to communicate with the insurer's customers. "Instead of looking at a phone call as a single transaction, we look at the information from the phone calls as a broader, business-intelligence perspective," WPS's Whitwam says.

WPS also allows call-center-agent recruits to listen to recordings of calls to give them a better idea of what to expect from their new jobs.

FedEx's Custom Critical unit, which provides exclusive transportation services aimed at customers with special handling needs, uses the software to help boost employee morale. When a customer uses the word "wow," or the emotion detection software detects gratification from a caller, the conversation is immediately recorded and can be shared with call-center agents.

What's next for emotion detection software? Artificial intelligence. Instead of users defining keywords and emotions, the software itself will figure things out, such as by analyzing voice pitch levels, a key determinant in emotion detection. By analyzing pitch, as well as tone, tempo, and inflection, the software in the not-too-distant future could be used to detect fraud. It already can differentiate between real anger and someone mimicking anger.

"We're taking art of call center management," Danon says, "and turning it into a science."

Licensing fees vary, but starts at $100,000 for up to 150 seats.

Never Miss a Beat: Get a snapshot of the issues affecting the IT industry straight to your inbox.

You May Also Like

More Insights