Blogs:
Rich Tehrani
Tom Keating
Greg Galitzine

Columnists: Tracey E. Schelmetic Charlotte Wolter David Sims Erik Linask Gary Kim Mae Kowalke Richard Grigonis Stefania Viscusi Tim Gray

By Mae Kowalke

TMCnet Associate Editor

 

Earlier this month, Microsoft announced that Speech Server 2007—the latest version of the company’s speech and telephony platform, which begins beta testing in May—will provide full support for VoiceXML.

 

VoiceXML is an open-standards-based language used for developing speech applications, managed by the World Wide Web Consortium (W3C).

 

In a press release, Microsoft (News - Alert) said the addition of VoiceXML compatibility (Speech Server continues to support Speech Application Language Tags, or SALT) will “enable customers to choose the development standard that will work best in their environment.”

 

According to Clegg Ivey, Vice President of Operations at Voxeo, Microsoft’s adoption of VoiceXML is a good indication that the language has matured into the industry standard.

 

“When Microsoft signs on to a standard, you know you're in good shape,” Ivey told TMCnet.

 

Since its founding in 1999, Ivey’s company has been a strong proponent of open standards for the speech industry. Voxeo built its platform on SIP, and used VoiceXML for its applications, before those protocols enjoyed industry-wide acceptance.

 

Back in those days, Ivey told TMCnet, many people in the industry thought that H.323, rather than SIP, would win out.

 

Luckily for Voxeo, today “SIP is well on its way to being as established as VoiceXML,” he said.

 

VoiceXML had an easier time becoming established than SIP, Ivey said, in part because it didn’t have as many non-proprietary competitors.

 

Another factor driving the acceptance of VoiceXML is the age of the speech industry itself, he noted.

 

Unlike with VoIP, Ivey said, “there have been speech-enabled applications since the 1980s.”

 

During the past several years, improvements to speech engines had a positive effect on the quality of applications such as text-to-speech and text-recognition, Ivey told TMCnet.

 

The result was an unleashing of pent-up demand for speech applications, which in turn helped the industry mature and accept that a universal standard was needed.

 

Until the late 1990s, when the W3C took charge of VoiceXML, there wasn’t an open standard for speech applications, Ivey noted.

 

But by 2001, he said, it became apparent that the speech industry as a whole would be better off if all players gathered around a single standard. Most agreed that developing an XML-based standard was the best choice.

 

Developing an industry standard, he said, was beneficial for customers, platform vendors, and developers alike. The only people who didn’t benefit were paid-by-the-hour gurus companies hired to help implement proprietary-based applications.

 

Today, because of the VoiceXML standard, companies can deploy speech products knowing that when they upgrade or switch vendors down the road, they won’t have to re-write all their applications.

 

Judith Markowitz, president of J. Markowitz Consultants, commented in an e-mail correspondence with TMCnet that “standards—especially VoiceXML—have transformed the speech industry and made speech technologies far more attractive to the marketplace than they were before.”

 

One result of VoiceXML’s acceptance is that vendors such as Voxeo now compete on factors like price and reliability, rather than proprietary languages.

 

“The one thing we don't do is use the different languages to trap customers,” Ivey stressed.

 

That’s good for customers, Markowitz noted, because “If you use accepted standards you reduce the risk of having to throw away everything you've done and start from scratch. This is true for any technology.”

 

The advantage of using an accepted standard hasn’t been lost on customers. Ivey noted that the vast majority of RFPs for speech applications now received by Voxeo include a requirement that the application be written in VoiceXML.

 

VoiceXML can be used to create any type of interactive speech application, Ivey told TMCnet. Some examples:

 

Voice-enabling of call centers. VoiceXML is ideal for building applications that understand what a caller is saying and respond in a natural-language fashion.

 

Interactive tracking applications. Examples are automated systems that allow you to trace the route of package shipments or the status of an airplane flight schedule.

 

Outbound notification. This type of application is used to send out automated calls which, for example, alert people of an approaching hurricane or remind them to refill their medical prescriptions.

 

“One of the keys to create a great interface is to think about what a natural-language conversation would be like, and replicate it,” Ivey said of these and other IVR applications. “You can do that very easily in VoiceXML.”

 

 

 

Subscribe FREE to all of TMC's monthly magazines. Click here now.