| Although numerous studies have indicated that | | | | received a three-year grant from the Air Force |
| customers have become more accepting of IVR | | | | Office of Scientific Research to investigate |
| systems, there are still instances that can cause | | | | emotion recognition and modeling in speech |
| frustration, especially systems that employ | | | | processing. Liu's research will focus on voice pitch, |
| sub-par voice recognition software. Frustrated and | | | | intonation patterns and word choice and their |
| angry users will tend to emote their feelings in | | | | associations with a variety of emotions such as |
| their speaking voice causing an already poor voice | | | | anger, sadness, happiness, surprise and frustration. |
| recognition system to have more trouble | | | | If successful, her research could be applied to |
| understanding the customer. This leads to a | | | | many applications including IVR systems that could |
| vicious cycle in which the user tends to become | | | | quickly recognize a frustrated user and |
| more frustrated while the IVR system becomes | | | | immediately transfer him or her to a human |
| less likely to recognize menu choices and/or voice | | | | operator. Such a breakthrough could result in less |
| commands. Ultimately this can lead to a very | | | | frustration among customers who are |
| negative customer experience which can effect a | | | | pre-disposed to dislike IVR systems. This would |
| company's bottom line. | | | | lead to a better overall customer experience and |
| Yang Liu, a computer scientist at the University of | | | | likely improve customer retention among |
| Texas at Dallas, is conducting research that could | | | | consumers who are resistant to automated voice |
| change the face of voice recognition software as | | | | response systems. |
| we know it. She and her graduate students have | | | | |