Microsoft Speech Server Announces Itself

    August 4, 2005
    WebProNews Staff

The product has been built on Windows and .Net; a future version of Exchange will be speech enabled.

The company lists Sandia National Labs as one of its new Microsoft Speech Server customers. Let’s hope someone at the weapons research lab doesn’t say “fire one” too close to a PC microphone.

Anyway, Microsoft has noted the debut of its speech platform in a press release from the SpeechTEK Conference. More than 60 companies will use the platform to launch speech-capable applications.

Our friends in Redmond say this is a “lower-cost scalable speech solution.” Besides Sandia, Microsoft lists King Pharmaceuticals and Virginia’s Department of Motor Vehicles as recent adopters.

Speech technology gives a call-center the opportunity to deliver various service options without human intervention. Fewer human agents means lower costs to the business. And if a speech platform can be acquired at a reasonable price, all the better.

Former Microsoft executive Kai-Fu Lee was an accomplished person in the speech recognition field, according to Microsoft’s biography on him. At Carnegie Mellon University, he developed the world’s first speaker-independent continuous speech-recognition system. He worked for Apple as VP of interactive media, which included PlainTalk speech technologies.

(And according to, Dr. Lee wrote a world champion Othello program while at CMU, taking a year off from writing a paper on speech recognition to do so. Then he went back and got his doctorate in speech recognition.)

It appears Microsoft’s speech group, at least, has moved on without Dr. Lee. The platform supports English, Canadian French, and US Spanish out of the box at no additional cost. And later, Microsoft plans to launch a version of Exchange with speech support.

According to the company, that effort will let Microsoft expand its speech offering “beyond the call center and interactive voice response markets into the broader enterprise unified messaging market.”

The platform supports Speech Application Language Tags and an ASP.NET server deliver it as a web application. Features include support for telephony-only, speech, or speech and visual applications.

Components of the Speech Server include the Telephony Application Services and the Speech Engine services, with automatic speech recognition capabilities. Microsoft expects further support for more telephony boards and VoIP applications from several of its business partners.

David Utter is a staff writer for WebProNews covering technology and business. Email him here.