When Exaudios is in use, it pays to get angry
New system can examine phone callers' intonation to tell they are happy, sad, in love, or angry. The technology is being pitched to call center managers.
PALM DESERT, Calif.--One of the coolest demos at DemoSpring here is from Exaudios.
The company's new Maginify product analyzes the intonation of someone on a phone call and can tell if the person is calm, beginning to get angry, or seconds from blowing up. It also identifies other emotions, but for this product, the focus is on happiness and anger. The service is being pitched to call centers, though there is potentially more to it than that.
You may rightly note that most humans have the innate ability to tell if the person they're speaking with is angry or happy, and that they shouldn't need software to tell them that. But simply knowing yourself if the person you're talking to is upset is not enough for call center managers. They'd like the capability to know which phone representatives are more likely to get callers angry, so they can correct their behaviors or let them see their own stats and self-correct. Also, using this service, callers can leave a message when they dial in, and the most upset of them can get routed to a live person more quickly.
So in the future, if you don't want to get stuck on hold, you may want to start yelling.
The product can also be used to collect data on reps, keeping track of how often they get angry themselves, and it can help route callers to the reps most appropriate to them. Some people are just better defusers of angry calls.
The Exaudios technology analyzes the intonation of spoken word. It is language-independent, CEO and inventor Yoram Levanon told me. In 10 seconds to 15 seconds, he said, it can gauge the emotional tenor of a voice on the phone, with about 80 percent accuracy. That's very impressive for a demo, but the false hits (identifying a person as angry when she's not, or vice versa) may make for a few awkward conversations and lost opportunities.
The intonation analysis software can also correlate with physical conditions. Levanon says it can be used to detect early cases of Parkinson's disease and schizophrenia, and that it can even detect heart disease. He also that the product is, for reasons he doesn't yet know, less accurate on people who have dyslexia.
I asked Levanon why he's focusing on call center technology when he could be working in medicine and improving lives. He said doctors are a conservative lot, and he's got to start the business somewhere. But if the technology proves out, I do hope that he pushes any diagnostic capabilities of the software more aggressively.
Maginify is available as an installed system for large call centers. There will also be a software-as-a-service version. At the lowest end, a version can be used to analyze recorded calls and generate statistical reports on them.