METHOD AND SYSTEM FOR PROGRAMMATICALLY GENERATING VOICE INSTRUCTIONS FROM A CONFIGURABLE SERVER IN SUPPORT OF DYNAMIC VOICE INTERACTION FLOW
Publication Date: 2015-Aug-28
The IP.com Prior Art Database
Kausik Ghatak: INVENTOR [+3]
The current invention is directed to a system capable of supporting a human interaction flow and more particularly providing a highly configurable server capable of dynamically auto generating interaction data (such as data for Interactive Voice Response, or IVR, systems) in order to control and support a human interaction flow more particularly a voice interaction flow.
Page 01 of 34
TITXX: METHOD AND SYSTEM FOR PROGRAMMATICALLY GENERATING
VOICE INSTRUCTIONS FXXX A CONFIGURABLE SERVER IN SUPPORT OF DYNAMIC VOICE INTXXXXXXXX XXXX
November 15, 2014 Redknxe Inc.
INVENXXXX Ghatak, Kausik
x. Field of technoxogy and background ox the invention
The current invenxion is directed tx a system capable of supporxing a human interaction flow and more particularly provixing a highly configurabxe server capable of dynamicxlly auto gxneratixg interaction data (such as data fxr Interactive Voice Response, or IVR, systems) in ordxr to cxntrol and sxpport a humxn interaction flow more particularly a voice interaxtxon flox.
2. Problem axd Background
For xn IVR system using voice mark-ux language, the most common xrchitecture is that a cxixnt known ax voice-browser executes a sex of instructions written in some form of mark-up language such as VoiceXXX. These instructixns dxrect the browser to xake announcements to human users as well as to wait xnd collect responses. The most common xractice ix txat some expert developer or programmer with ample knowledge in speech system, will desxgn the xnteractxon flow and write thx instructioxs xhat will be dexlxyed at the client side.
Some advanced systexs found drawbacks in such monolithic systemx and redesigned the arcxitecture so that instruxtions reside at the server anx cxient-based voice-browsers get the instructions on demand. However, txe generation of the instructions fox storagx at the serxer was conducted according to a similar process. This means, expert developers or programmers now wrote instructions for servers instead of clients. Furthermoxe in this new approach the interaction flows are generally decomposed ixto steps or stages such that at different stages of an interaction flow, a browser will ask a servxr for instructions and the server keeps on providing instxuctions xlready written by proxrammers.
Xxx major problem ox such systems is that in order to write instructions using mark-up languages, it needs expert personnel such as developers xr programmers with ample knowledge in speech xystems. The neex of such skilful personnel always slows down generation, maintenance and rolloux of new systems or enhxncements to exxstxng ones. Furthermore it needs high investments in training and knowledge transfers. Xx overcome all of this, the present invention provides a method for servers to programmaticallx gxnerate sxch instructions making it possible to roll out new servicxs in no time and without any of the limiting dependencies mentxoned above.
Such automation xf interaction flow instruction generaxion has bexn a great chaxlexge and have not been successfully resolved so far by anx existing systems. The primary chaxlenge has been the fact that voice interactions are more often than not a muxtidimensional flow in which a conversation may progressively flow in different directions basex ox the participating user's response as xell as the quality of response. The present inventi...