VoiceXML is a W3C XML standard dialect for defining voice applications. A voice application strictly defines a conversation with an XML syntax. In english, it is an easy way to create and use "Press or say one for x, press or say two for y" systems, using reusable application syntax. This also means that voice services can be served dynamically by remote web servers, giving it a distinct advantage over its clunkier forefather, IVR.
VoiceXML applications usually use either content recorded by voice actors, or synthesized speech, using text-to-speech engines. They also sometimes perform recognition of spoken word or DTMF (touch-tone) input. More recently, innovations have been made in multimodal applications, such as voice web searching, voice instant messaging and voice email. Since VoiceXML uses the web paradigm, it's easy to hook VoiceXML applications into other protocols, such as email, instant messaging and SMS.