Lucent Technologies Inc. is taking advantage of the recently-developed VoiceXML specification and is deploying a speech server.
VoiceXML is an open programming language used to make Web content and services accessible by voice commands via telephones. Version 1.0 of the specification was released last month by the VoiceXML Forum, which is also pushing for standardization.
Lucent’s Speech Server is a network-based server that lets service providers offer voice access to the Web by speaking into a phone.
The server is capable of matching a phone number with specific URLs, the content of which is then read out to the caller in voices created using Bell Labs’ text-to-speech software.
The Speech Server can scale to support millions of users, according to Lucent. Because it was built using open interfaces, the server can be used with any other vendor’s equipment and can be placed in voice or data networks, Lucent officials said.
The VoiceXML specification is backed by more than 100 companies, including Cisco Systems, Ericsson, Hewlett-Packard, Nortel and Novell.
Using VoiceXML, application developers can tailor specific applications for use by any VoiceXML-compliant product, according to its proponents.
One such developer, a company called VoiceGenie.com, will release at Internet World its VoicePortal+, which is based on VoiceXML technology.
VoiceGenie’s VoicePortal+ is a content aggregator accessible by wireless, wireline or cable phone connections, according to the company. VoicePortal+ is aimed at Web portal companies such as Netscape, Yahoo and Excite.
According to one analyst, VoiceXML holds huge potential but has issues to overcome.
“Everybody agrees on the potential, but [VoiceXML] hasn’t been adopted as a standard yet,” said Bill Meisel, president of TMA Associates in Tarzana, Calif. “It is a little vague in some places on specific requirements, so some browsers that use it will have proprietary limitations.”
Meisel added that the technology may force companies to rethink their Web sites because “Web sites designed to be seen cannot be easily understood in audio.”