A brief introduction to VoiceXML

I think that one of the quickest ways of describing VoiceXML is to say that it's IVR XML. It allows you to write open standard IVR applications using web technology.

So what?, I hear you say. Well the what is NO VENDOR LOCK-IN. In the good old days you bought an IVR and there was little you could do with it without your supplier getting involved. Not only did you buy the IVR from them, you also bought the applications and their maintenance from them as well. Want to change a prompt? Call your IVR supplier. Shouldn't take more than a couple of weeks and they would charge accordingly as well.

These traditional black-box IVRs were all-in-ones - they had two connections to the outside world: the power cable and the telephony cable. They were designed for linear DTMF driven IVR hell. And they did their job well - but just try writing a mixed-initiative ASR application with one sometime.

But VoiceXML doesn't just make you flexible in terms of your supplier, it goes beyond that.

VoiceXML is specifically designed for writing DMTF and ASR user interfaces - it takes care of a lot of the drudge work like tracking how often the caller has said nothing, or said something that wasn't recongised.
VoiceXML uses the web client/server approach to application development - this means there's a large developer community out there ready to write applications.
As well as that it allows you to add a voice channel to existing web services, not to mention being able to reuse your existing web server infrastructure.
VoiceXML doesn't specifiy which TTS engine to use - nor does it specify which ASR engine you use. All of the interactions with the technologies are abstracted.
Nor does it specify whether you have to use PHP, Perl, Java, JSPs, ASPs, etc. to create the VoiceXML pages.
VoiceXML allows you to consolidate your IVR infrastructure. There is now no need to have separate IVRs often from different suppliers for different services. You can now migrate all services to VoiceXML and use just one IVR.
This consolidation of infrastructure allows you to lower OPEX, as you are no longer maintaining several IVRs.
VoiceXML is so cool, you don't even need your own IVR! Just rent a few lines from a hosting company, and write your own VoiceXML applications. It's the web - your applications are requested using http and rendered by the VoiceXML browser.

Another important point is that when you write a VoiceXML application, your investment is better protected. In theory your VoiceXML application should run on any VoiceXML browser. One or two minor modification may be necessary - but the core application and data remain the same. All that has to be changed is the formatting so that it can be rendered by the other browser. With a proprietory IVR there isn't much that can be migrated - change platforms then you write the application again from scratch.

VoiceXML is just a mark-up language and a VoiceXML browser is just that - a browser that renders VoiceXML acoustically. And of course it's cool!

If you have any comments, ideas, issues, etc. about this topic why not try the voice-push forums