In an effort to move computing to smaller devices, such as hand-held computers and cellular phones, IBM is revamping its voice-recognition technology. The company, which has owned a stake in the speech business for the past 30 years, is shifting its focus to delivering core technologies, rather than individual products. As a result, IBM will gradually introduce seven new products for companies to enable speech recognition in their products.
The first technology to bring this strategy to the forefront is arriving this fall. The Web Sphere Voice Server, a Web-based infrastructure platform that will provide voice-enabling e-business options, uses ViaVoice technology that runs on a VoIP platform and a direct talk platform. This product supports VoiceXML and Wireless Application Protocol, as well as VoIP.
Embedded ViaVoice is the backbone of a plan to get embedded voice-recognition technology into as many handhelds as possible.Ozzie Osborne, general manager, IBM voice systems, says by using this voice server people can access Web pages by calling from a phone and punching a code. The server provides a platform for voice products to access information 24/7, using the same logic as data and business logic.
"The last few years, we have been driving it toward the enterprise as a technology," Osborne says. "What we're trying to do is build an end-to-end distributed platform and tools for people to create speech applications." Osborne predicts that in three to four years, people will be able to transact business and get information anywhere, on a PDA or phone, at any time of the day.
"Right where we are (with voice technology) is where the Web was in '93," Osborne says. "In the next three years, there will be a massive explosion in voice applications."