Mobile // Mobile Applications
Commentary
8/14/2006
04:06 PM
Commentary
Commentary
Commentary
50%
50%

Don't Ask, Don't Tell...Your Computer

Perhaps it will be a stroke of genius to link voice recognition into established IT apps, or perhaps it will be one more piece of bloatware that IT doesn't really want to have to deal with.

After over 20 years in No Man's Land, it seems like speech recognition is finally finding its groove: telephone-based customer service.It's certainly true that speech recognition has a devoted user base, especially among differently abled workers. The technology has long helped those who aren't able to see their computer screens well or at all, and so rely on speech recognition to read a Web site for them, or help input a word processing document. It has also proved useful for people with repetitive stress injuries and other people unable to use a keyboard and mouse.

As much of a boon as it's been to the disabled community and to their co-workers, customers, and partners, who get the benefit of their skills, speech recognition has never really achieved the type of mass-market promise that manufacturers have, well, talked about. The last time I wrote anything in-depth about speech recognition was in 1984, for InformationWeek when I worked here as a cub reporter just learning about the tech industry. (Sorry, I can't give you a Web link for something that far back; you're just going to have to trust me on this one.)

Back then, the dream was of cubicles full of white-collar workers all talking into their computers--in other words, many, if not all of us, would use speech as the user interface of choice in the not-too-distant future.

Turns out that scene from The Jetsons didn't come to pass for technical as well as practical reasons. First, the software typically needs a training period to get used to a given person's speech patterns and cadence. Any two people, even of the same generation and from the same geographic region, will pronounce words a bit differently, and the software needs its owner to read through fairly hefty a list of words, at least twice, to understand when you say "girl" versus "pearl."

In fairness, I must point out that our most recent review of Dragon NaturallySpeaking found it does reduce the amount of training required. Another colleague agreed this package has made strides.

But even once trained, most speech software simply isn't sophisticated enough to differentiate among more than one cubicle mates talking, especially at the same time, even if one is in the background. Imagine the confusion with dozens of people in cubicle farms talking to their PCs simultaneously.

Third, the accuracy rate hasn't been that great, as anyone who's wrestled with "customer service" telephone systems can attest. My personal favorite is when you call the main switchboard number and get one of those verbal employee directories. Nine times out of 10, I get the wrong person and have to call back three times to figure out how to reach an operator. (System: "Please speak the name of the employee you wish to talk to." Me: "Tony Danza." System: "You selected John Smith. Please hold while I connect you.")

I can understand if I'm mistakenly connected with a person whose name sounds similar to the one I may have mumbled, or if they're at least close to each other alphabetically. But most of the time, the name of the person I get connected to has neither of those characteristics.

It's one of the mysteries of life, along with where the socks go when you know you've put a matched pair into the dryer and only one comes out again.

The relatively few times speech-driven telephony has worked for me have been perfectly efficient, I must say. But it just doesn't work too often. So given this experience, I read with some skepticism about how Microsoft and IBM are now competing to tie voice recognition in with their other applications--Office, in Microsoft's case, and WebSphere, in IBM's.

Perhaps it will be a stroke of genius to link voice recognition into established IT apps, or perhaps it will be one more piece of bloatware that IT doesn't really want to have to deal with.

In either case, Microsoft's recent embarrassment with its speech demo in front of a roomful of financial analysts seems to prove my point. The system wrote "Dear Aunt" when the presenter said, "Dear Mom."

And one other note of interest: Nuance, which makes Dragon NaturallySpeaking, and Microsoft both recently promised to work with the GetHuman project, which has established a database of secret phone numbers and codes that enable callers to skip directly to a human when they call customer service.

What do you think? Is there a place for speech recognition in today's customer service organizations, and how do you feel about supporting it from an IT standpoint? Please comment below.Perhaps it will be a stroke of genius to link voice recognition into established IT apps, or perhaps it will be one more piece of bloatware that IT doesn't really want to have to deal with.

Comment  | 
Print  | 
More Insights
Building A Mobile Business Mindset
Building A Mobile Business Mindset
Among 688 respondents, 46% have deployed mobile apps, with an additional 24% planning to in the next year. Soon all apps will look like mobile apps and it's past time for those with no plans to get cracking.
Register for InformationWeek Newsletters
White Papers
Current Issue
InformationWeek Tech Digest, Nov. 10, 2014
Just 30% of respondents to our new survey say their companies are very or extremely effective at identifying critical data and analyzing it to make decisions, down from 42% in 2013. What gives?
Video
Slideshows
Twitter Feed
InformationWeek Radio
Archived InformationWeek Radio
Join us for a roundup of the top stories on InformationWeek.com for the week of November 16, 2014.
Sponsored Live Streaming Video
Everything You've Been Told About Mobility Is Wrong
Attend this video symposium with Sean Wisdom, Global Director of Mobility Solutions, and learn about how you can harness powerful new products to mobilize your business potential.