Phone call emotion analyzer is a winner (w/ Video)

( -- An Israeli company, eXaudios Technologies, has developed a software application that automatically analyzes the intonation and volume of a person’s speech in real time and translates it into statements and graphics describing the speaker’s emotions.

The system, dubbed Maginify, is likely to find its way into call centers to monitor calls, where it will allow the staff to identify callers who are becoming angry, and react to appease rather than inflate the . It will also enable supervisors to monitor the agents, and even step in if required, to try to turn the call around. Managers will be able to identify agents who are behaving in a way that tends to anger customers, and agents may be able to adjust their own behavior when they see their call statistics. Callers will also benefit because they will be able to leave a message when they call in, and the most upset or angry would be pushed to the front of the queue to speak to a live agent.

The system identifies a range of emotions, but in the call center environment anger and are the most important in terms of keeping customers, and the profits that often come with them. The speech analysis technology can also correlate with physical indicators, which according to eXaudios could one day make it useful for screening for Parkinson’s, schizophrenia, and even heart disease. eXaudios also believes its product could find application in in the US.

DEMOSpring 2010 video.

The Maginify system is independent of the language spoken and analyzes the of the voice. Its inventor and CEO of eXaudios Technologies, Yoram Levanon, said the system produces a statement of the speaker’s emotions in 10-15 seconds, with around 80% accuracy. Levanon said that for unknown reasons the technology is less accurate when the speaker has dyslexia.

The first use of the technology will be installation in large call centers, but it will also be available as a software-as-a-service release. There will also be a low-end version that analyzes recorded rather than live calls and generates statistical analyses on them.

The system was unveiled at DEMOSpring 2010 in Palm Springs, California earlier this week, where it won the People’s Choice Award of advertising to the tune of $1 million.

Explore further

Spy software used in call centers

More information: Company website:

© 2010

Citation: Phone call emotion analyzer is a winner (w/ Video) (2010, March 25) retrieved 20 September 2019 from
This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no part may be reproduced without the written permission. The content is provided for information purposes only.

Feedback to editors

User comments

Mar 25, 2010
How does this device address differences in cultural patterns?

For example, A french Canadian speaking english will typically sound as though he's asking a question when making a statement, while an American will sound like he's making a statement while asking a question by intonation.

Likewise with Chinese and Japanese, the intonation is wholly different with Chinese starting loud and tapering off and Japanese doing the opposite.

Interesting technology though. Potential uses are extraordinary.

Mar 25, 2010
"where it will allow the staff to identify callers who are becoming angry"

Steps to get to an operator quickly:

System: "The wait is currently 15 minutes. State the department you are trying to reach".
You: "Customer support dammit!!".
Operator: "How may I help you".

You'll never have to wait again. Enjoy!

On a serious note, I agree with Skeptic Heretic. Many times it is hard to discern tone as a human being. I can image that the technology will have to be custom calibrated to the region of use. Or perhaps it will recognize accents and adjust its parameters.

Mar 25, 2010
All I know is this better work better than those lie or love detectors based on voice analysis that came out of Israel a few years back - they were total junk...

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more