Speech Server 2007, the latest version of Microsoft's product for building interactive voice response (IVR) applications, will include support for VoiceXML, a widely used markup language for IVR applications, as well as improved integration with Visual Studio and the .NET Framework and support for Voice over IP (VoIP). A beta of the product will be available in May 2006.

Speech Server is a platform for building IVR applications, which allow users to make choices by speaking or pressing numbers on a telephone keypad and retrieve information in the form of prerecorded messages and prompts or computer-generated speech. IVR applications are particularly valuable in automating customer service call centers.

VoiceXML, .NET, and VoIP Support

Current versions of Speech Server allow developers to create ASP.NET Web pages that use the Speech Application Language Tags (SALT) to identify how a user interacts with the application via voice. For example, a field for the customer's account number could be marked with a set of <prompt> and <listen> tags that define the voice prompts provided by the application (such as "Please state your account number"), list valid inputs it accepts, and specify how tones from the telephone keypad should be interpreted. By using ASP.NET, Speech Server also helps developers build applications that can be accessed by voice, by Web, or by a combination of the two (so-called multimodal applications).

In addition to SALT, Speech Server 2007 will support VoiceXML, an alternate markup language developed by a consortium of companies including AT&T, Lucent Technologies, and Motorola, among others. Speech Server 2007 will include a VoiceXML editor that integrates with Visual Studio 2005. VoiceXML and SALT are similar but have different histories. VoiceXML was developed in 1999 for creating traditional IVR applications while SALT was developed in 2002 for voice-enabled applications for the Web and devices such as personal digital assistants (PDAs). The new support could make it easier for customers to switch to Speech Server 2007 from competing products based on VoiceXML.

Speech Server 2007 will also provide .NET APIs that will allow developers to build speech applications without ASP.NET, to entice developers who are more comfortable with traditional-style APIs in addition to the Web developers Microsoft has courted with previous versions.

Speech Server 2007 also provides support for several important VoIP standards, including the Session Initiation Protocol (SIP) and Real-Time Transport Protocol (RTP). With this support, Speech Server can communicate with VoIP call-center equipment, which many organizations are using to replace circuit-switched telephone networks with less expensive Internet-based networks.

Reporting and Analytics

Speech Server 2007 will also include a new Analytics Studio, built on SQL Server 2005 Reporting Services, with predefined reports that could help call-center managers better track the efficiency of their operations. The Analytics Studio also allows managers to drill down from aggregate statistics to details on individual calls, including listening to the call itself.

In addition to the improved reporting, Speech Server 2007 will provide a set of business intelligence tools for managing call data from long historical periods.

Resources

Details on Speech Server 2004 can be found in "Server to Aid Voice Response Applications" on page 12 of the Jan. 2004 Update and "Speech Server for Voice Applications" on page 17 of the May 2004 Update.

The Speech Server 2007 home page is www.microsoft.com/speech/default.mspx.

The VoiceXML Web site is www.voicexml.org