The Monitor

TDVoice recognition is a 64-year-old technology that is finally breaking ground. For the first time since its inception, voice recognition is up to the quality necessary to become useful and ubiquitous. It's a technology that sales and marketing executives must watch: The emerging speech recognition products and services will bring voice access to Web sites, essentially enabling customers to phone sites. In fact, The Kelsey Group predicts that there will be 3 million frequent and 10 million occasional users of voice-enabled Web sites by 2002, and 30 percent of all new IVR telephone ports will be speech enabled by 2003, according to GartnerGroup.

This easy-to-use technology offers myriad benefits, including providing users significant time savings and productivity improvements. Voice recognition can be multilingual or voice-specific. It takes only minutes to learn and its computerless connection to the Internet opens up the Web to a whole new set of users. Of course, every emerging technology has problems as well. And voice recognition does have its share—though none so heady as to deny the value of the benefits. The primary obstacle to voice recognition's growing ubiquity is user acceptibility: Will people speak as freely and naturally into a kiosk or a watch as they do into a cell phone? Another challenge is that the emerging voice XML standard is not yet the standard of choice, nor is it obvious how or when it or another technology will be standard. And ever-popular handheld computers don't yet offer compatibility with some voice-recognition software.

Even with the percentage of recognition (how much speech the software can understand) now in the high nineties, the opportunity is outrageous. Add voice recognition with wireless, and there's no more plugging in to sync computers. Give salespeople, who typically hate to type, voice recognition and speed the delivery of reports and proposals.

Industry leaders IBM and Lernout & Hauspie are pouring big money into development; new players like Conita, Periphonics, Speechworks, and phone.com are offering creative uses of the technology. Charles Schwab and Fidelity are using voice recognition to place trades, which saves large sums of money; Duke University Medical Center's radiology department used it to decrease the time it takes to process an x-ray from 48 hours to 3 hours. Oracle claims that its portal will soon be voice enabled: Talk to a page and the page talks back. It's a big wave coming.