Plum IVR Support

The Plum Voice Platform supports version 2.0 of VoiceXML.

Any executable VoiceXML code can go within a <catch> block. However it should be noted that great care be taken when catching a telephone.disconnect.* event: If a <catch> block contains VoiceXML code that generates another disconnect event (for instance a <prompt> within a <catch> block will throw a...

Only prompts within an executable block can be interrupted by a disconnect event. All other executable code will be run to completion.

The <param> tag submits data to a <subdialog> (or <object>) without the need for an HTTP request or server-side scripting. Any data passed via <param> is available as a series of variables in the (new) <subdialog> context. E.g.: <form id="test"> <var name="var1" expr="'tooth'" /> <var name="var2" ex...

Yes. If you specify a <grammar> element within a <choice>, that grammar will be used instead of the automatically generated grammar. <menu> <prompt>Say sports, weather, or news</prompt> <choice next="#sports"> <grammar>sports</grammar> One </choice> <choice next="#weather"> <grammar>weather</grammar...

Input items (<field> tags, <record> tags, etc.) have an associated set of shadow variables which are used to return results from the execution of the input item (other than the value stored under the input item's name attribute). A shadow variable is referenced as name$.shadowvar where name is the v...

This depends on the active input modes and the bargein attributes of any prompts or blocks that precede a given input item. Active input modes can be set at a given scope (dialog, document, etc.) by setting the global platform property "inputmodes". <property name="inputmodes" value="dtmf voice" /> ...

You can spell out any arbitrary string of text by wrapping a <say-as type="acronym"> tag around it. E.g. <form id="sayas_test"> <block> <prompt> <say-as type="acronym">MSFT</say-as> is currently trading at <say-as type="currency">$25.03</say-as>. <say-as type="acronym">CSCO</say-as> is currently tra...

The Bargein Threshold and Silence Threshold can be set from the administrator's Server Configuration form. The values set there are in decibels where 3 dB represents the loudest possible signal over the phone. The bargein threshold is specifically the loudness level that must be reached before the p...

The <prompt> tag's "count" attribute specifies how many times the user must have visited the form item containing a given prompt before it is played. E.g. <form id="form1"> <field name="blah" type="digits"> <prompt count="1"> Please enter a number. </prompt> <prompt count="3"> Using your telephone k...

The # key is the default key for terminating DTMF input. The # key itself can be captured if it is specified within a grammar: <form id="pound_test"> <field name="foo"> <grammar>1|2|"#"</grammar> <prompt>Please press one, two or the pound key</prompt> <filled> <if cond="foo == '#'"> You pressed the ...

Yes. In the Application Setup window next to each DNIS in your account there are three links: Last, Live, and Utterance. If you click on the Utterance link you'll have access to .wav files of all of the utterances for the DNIS from the last call session.

There are two possible recognition modes: speech and hotword . Speech mode is the only recognition mode support by the Plum Voice Platform. In speech mode, the platform will interrupt a prompt and attempt recognition as soon as it detects any sound. The platform will then analyze the sound for speec...

The Plum Voice Portal supports multiple speech recognition engines, and thus many grammar formats. The two main grammar formats are SRGS and JSGF. In both cases, substrings of the utterance are actively searched for. So, in SRGS: <grammar type="application/srgs"> <rule id="action" scope="public"> <o...

No. The Plum Voice Platform supports state of the art automatic speech recognition. However, the state of the art -- from Plum or any vendor -- does not support reliable, speaker independent transcription of arbitrary speech. Some speech recognition systems are tuned to a particular speaker and requ...