Slashdot Mirror
System Recognizes Emotions In People's Voices
cylonlover writes "Automated telephone services may get slightly less annoying thanks to research being carried out at Spain's Universidad Carlos III de Madrid and Universidad de Granada. A team of scientists from those institutions has created a computer system that is able to recognize the emotional state of a person speaking to it, so that it can alter its behavior to make things less stressful."
That could become the ultimate trolling device!
Why not start out with the less stressful option, just saying...
Caller: Oh it's another damn machine.
IVR: Do not take that tone, please, sir or madam.
Caller: WHAT tone?
IVR: Please remain calm, and speak clearly.
Caller: I AM CALM, DAMN IT!!!
Admit it. You post strawman arguments as AC so you get modded Insightful for refuting them, rather than Troll
just have the system scan for curse words. easier than detecting emotion, and probably more accurate to determine when people are hating the IVR.
Source: HAL: Look Dave, I can see you're really upset about this. I honestly think you ought to sit down calmly, take a stress pill, and think things over.
Shh.
... from the Sirius Cybernetics Corporation.
Sara
Designer, Gamer, Macgrrl in an XP World
If (person_calling)
connect_to(actual_human);
Guns don't kill people; Physics kills people! - John Lithgow as Dick Solomon on Third Rock From The Sun
have a competent human answer the fucking phone.
Shoes for Industry. Shoes for the Dead.
Is this one of those "On this day in history..." stories? Because analyzing caller voice patterns for stress has been SOP for the big boys for years. A pretty common way to get out of the voice prompts and to a person who will likely be competent enough to help you is to swear profusely at the voice recognition system. You'll then often be passed to a senior CSR who can get shit done. The catch is that they're authorized to hang up on you more quickly than a regular rep if you continue to swear once they're on the line.
I guess the news here is that the existing technology is being used to present automated scripts tailored to the individual caller.
How about just using the low-stress behavior ALL the time?
Because "don't stress out the human" might be actually really unhelpful stuff like slowing down speech, presenting fewer options at a time, or even fewer options at all.
Computer: "Sears tool desk, can I HELP you?"
Me: "Operator."
Computer: "I'm sorry, I didn't understand that. Did you say, screwdrivers?"
Me: "Salesman"
Computer: "I'm sorry, I didn't understand that. Did you say, salepaper?"
Me: "Cashier."
Computer: "I'm sorry, I didn't understand that. Did you say, chainsaw?"
Me: "NOW I want a chainsaw! I'm coming down there and #!*(%$!*%^(!"
Computer: "I understood that! Calling 911."
"He who would learn astronomy, and other recondite arts, let him go elsewhere. " -- John Calvin, commenting on Genesis 1
Given the widespread availability of the Internet, the telephone is really quite a useless device now, especially for business purposes.
When dealing with any business or institution, it is often much easier and much more effective to use a web app of some sort. These kind of apps work perfectly fine for most people, and they avoid a lot of the confusion and annoyance that can happen when having to use a telephone-based system. In the remote chance that the user needs to deal with a live person in realtime, there are many chat systems available that work just fine.
Spoken communication is a relic of the past. It is usually no faster than using a web app or other written communication, and most of the time is much slower and much more prone to inaccuracies creeping in. In the vast majority of cases, the phone shouldn't even be an option. If it is, it should merely be the last resort.
If that were accurate, those giant call centers that companies employ in addition to their websites would go largely unused. Turns out, lots of folks prefer to talk to someone. Just because you don't does not mean that it's a "relic of the past". While I like to do business on the web, I will most definitely avoid shopping with a company if I can't find telephone contact information easily. Web stuff works great when things are going smoothly, but typically the moment you have a hiccup in the standard process, you need to get someone on the phone to fix it. IVRU's are just a way to waste your time while waiting to be helped, so I see this effort as a mostly empty gesture. As GP said, why not start out with the less stressful option in the first place.
Unless someone can make voice recognition actually works no amount of behavior altering is going to help. Maybe it's just me, but it doesn't seem like voice recognition software has come a whole lot farther than where it was when I first started playing with it over a decade ago.
giggity
yes, because all official websites are well designed, and provide many options outside the normal list that can tailor to the needs of every single user.
sometimes a second brain is needed in a transaction.
I intentionally do business with companies that don't employ long, painful, circular IVR. I simply won't do business with a site that doesn't have a phone number.
As you said, the web is great if you're trying to buy something. Beyond that, web interaction is mostly garbage. Many companies don't provide direct email support anymore. If they do, it usually takes 3 days to hear back because email is a nice... leisurely... game... of... ping... pong... that... they... answer... when... it's... convenient... for... them.
Those horrible IVR's are used to deflect you back to the web. Not because their web resources are better than their people, but because humans on the phone are an expensive and time consuming resource they don't want you to use.
So it really is a place to separate yourself from the rest. If I call your number and actually get a person that can fix problems, you're head and shoulders above your competition.
Real time speech analytics for call centers has existed for seven years. The better products came out of Israel, at least the first and second generation IP Telephony capable systems. Inflection based triggers have existed in traditional TDM systems for over twelve years, so not real sure why this specific article is so intriguing. Just about any high end inbound call center will use some form of inflection and emotion algorithmic processing, more so once you get into the arm pit of finance, collections.
The systems I manage process over 500 million calls per month, across multiple industries and pbx vendors. The majority of the volume is processed using Cisco and Interactive Intelligence products, with an assortment of one off custom solutions. Of that 500 million calls per month, over 30% of the calls have used some form of inflection and emotional detection within the last eight years.
*yawn* another slow news day?
When the only tool you have is a hammer, every problem looks like a nail
Mimicking the response of understanding, to manipulate other's behavior.
You are in a twisty little maze of voice options, all alike.
"Flyin' in just a sweet place,
Never been known to fail..."