Microsoft will release a beta of the next version of its speech and telephony server software in May, the company said Wednesday.

Microsoft plans to release a final version of Speech Server 2007, its second major release of the product, in late 2006. Microsoft launched Speech Server, which allows customers to deploy speech-recognition and intelligent voice-response systems, in 2004 and released an interim update last year. Speech Server is still in its early-adopter phase, with about 100 customers in Canada and the United States.

Speech Server 2007 will focus not only on making it easier for developers to build speech applications using the software, but also on ensuring those applications provide a satisfying user experience, said Clint Patterson, director of product management for Speech Server at Microsoft.

While speech-recognition and interactive voice response systems have been in use for some time—mainly as systems to handle customer service telephone calls—they still remain largely unpopular with end users.

"As a consumer, I have an aversion to these applications," Patterson acknowledged. "But our mission is to make customers consider these applications like [they do] an ATM or airport self-service kiosk—it’s a convenience to have these things because you get your task completed more quickly or efficiently."

To that end, Microsoft plans to include better application prototyping tools in Speech Server 2007, as well as tools for diagnosing how speech applications might be succeeding or failing with customers so the user experience can be improved.

Tools to achieve the latter include Speech Server Analytics Studio and Speech Server Business Intelligence, which will give companies detailed usage reports of how customers react and interact with speech applications, he said.

Another key new feature of Speech Server 2007 is native voice-over Internet protocol (VoIP) support, Patterson said. Previously, customers needed third-party telephony hardware to deploy an application built using Speech Server on a VoIP system, he said.

Microsoft also is including support for a development standard for voice applications called VoiceXML (voice extensible markup language) in Speech Server 2007, Patterson said. Developers building voice recognition and speech applications either use SALT (speech application language tags) or VoiceXML as part of the development of those programs, he said. Previously, Microsoft only supported SALT natively in Speech Server.

All of Speech Server 2007’s new features will be available as part of next month’s beta with the exception of Business Intelligence, which will be in the product once it is released to manufacturing, Patterson said.

Customers interested in signing up for the beta, which will have limited participation, can visit the company’s website.