Taiwanese researchers have developed an automated technology that can gauge customer satisfaction over the phone. And it can do it better than humans.
Shuchih Ernest Chang and Yu-Teng Jang of the National Chung Hsing University in Taiwan used recorded customer voice files to “construct an artificial neural network-based” technology that matches tone of voice to emotions and then matches emotions to customer satisfaction levels.
As described in Assessing Customer Satisfaction in a V-Commerce Environment, the researchers developed a database of vocal tones associated with satisfaction and happiness levels. (Here’s a v-commerce definition, by the way.)
From the report:
Prior studies found that emotion can be inferred/extracted based on certain features of speech, and emotion also influences satisfaction. From an integrated viewpoint, our research finding is consistent with the findings from these studies, suggesting that satisfaction is associated with emotion and emotion is associated with speech, respectively.
By analyzing variables in sound-wave patterns in a customer’s voice, the technology assesses tone and matches tone with emotions in the database.
Chang and Jang recorded customer-service calls into an IVR system. Afterwards, they had the technology and human counterparts analyze the calls. They then compared the analyses to post-call surveys the customers completed.
The technology was closer to the survey results than the humans were. It scored 83% accuracy, whereas the humans only managed 62.5%.
The researchers believe the technology is a viable alternative to questionnaires, which many customers are unwilling to complete. They assert that it can determine customer satisfaction through audio, in real time. Not only that, it can save the data to help organizations better adapt the system to their unique callers.
The proposed method and system can detect customer satisfaction on the spot; i.e. it detects customer satisfaction in real time. Once we obtain the customer satisfaction level in a real-time v-commerce environment, timely feedback can be given to those relatively unhappy or unsatisfied customers for achieving the goal of timely service recovery.
Chang and Jang see the system working in “various voice-based business applications, such as call centers and customer relationship management, to achieve the business objective of improving customer satisfaction, enforcing customer loyalty, increasing re-purchase rate and enhancing enterprise’s benefits.”
If the technology is viable, we could theoretically develop automated voice systems that could identify, in real time, when a customer is unhappy, and then act on the information.
A customer calls in, isn’t happy. The IVR notifies a live agent or customer experience manager, who addresses the problem themselves via phone, social media or however.
And/or we have the technology act as well. It could follow up with a kind or helpful voicemail, send a special offer email, text an invitation for further discussion, et cetera.
If viable, the technology could help us all achieve the high-touch service we want to provide.