METHOD AND SYSTEM FOR VOICE ORDERING UTILIZING PRODUCT INFORMATION

• US 20090319272A1
• Filed: 06/30/2008
• Published: 12/24/2009
• Est. Priority Date: 06/18/2008
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: International Business Machines Corporation
• Inventors: Occello, Audrey, Palaysi, Christophe, Coulomb, Beatrice, Gibault, Dimitris
• Granted Patent: US 8,321,277 B2
• Time in Patent Office: Days
• US Class Current: 704/251
• CPC Class Codes: G06Q 30/0603 Catalogue ordering
• CPC Class Codes: G10L 13/00 Speech synthesis; Text to speech systems

Abstract

A method for voice ordering utilizing catalog taxonomies and hierarchical categorization relationships in product information management (PIM) systems includes: prompting a user with a query to input speech into a speech recognition engine; translating the inputted speech into a series of words; querying a product information management component (PIM) based on the series of words; wherein the querying is performed as a matching algorithm against PIM category and attribute keywords; returning coded results to a voice synthesizer to produce at least one of: a voice response, and a text response to the user; and wherein in the coded results indicate one or more of: a not found message for zero matches, a confirmation of a suitable single match, a request for additional information in the event one or more of the following occurs: more than one matching item, category, and item attribute was found in the PIM.

Claims

1. A method for improved voice ordering utilizing catalog taxonomies and hierarchical categorization relationships in product information management (PIM) systems, the method comprising:

prompting a user with a first query to input speech into a speech recognition engine;

receiving the user'"'"'s inputted speech;

translating the inputted speech into a series of words;

querying a product information management component (PIM) based on the series of words;

wherein the querying is performed as a matching algorithm against PIM category and attribute keywords;

returning coded results to a voice synthesizer to produce at least one of;

a voice response, and a text response to the user; and

wherein in the coded results indicate one or more of;

a not found message for zero matches for an item in the PIM, a confirmation of a suitable single match of an item found in the PIM, a request for additional information from the user in the event one or more of the following occurs;

more than one matching item, category, and item attribute was found in the PIM.

2. The method of claim 1, wherein items, categories, and sub categories are organized in a hierarchical structure of levels;

wherein each level has its own attributes;

wherein the keywords follow the same hierarchal structure as in the PIM being searched; and

wherein relationships between the keywords are also taken into account during the search of the PIM.

3. The method of claim 1, wherein in the event there are no identified item matches based on the first query, a determination is made if there are any category matches;

wherein in the event there are no category matches, an error message is generated by the system processor and sent to the voice synthesizer, to generate an additional request for voice inputs into the speech recognition engine; and

wherein in the event there are category matches, a request is generated by the system processor for further sub category or item attributes via the speech recognition engine.

4. The method of claim 1, wherein in the event there is one potential identified item match, the item'"'"'s code is passed by the system processor to the voice synthesizer, to generate a request for the user'"'"'s confirmation of the matched item;

wherein in the event the identified matched item is incorrect, a request is generated by the system processor for further sub category or item attributes via the speech recognition engine, while keeping a record of the current results to be refined at a subsequent interaction with the ordering user; and

wherein in the event the identified item is correct, the ordering process of the identified matched item proceeds.

5. The method of claim 1, wherein in the event there is more than one potential identified item match, a determination is made if there are category matches corresponding to the item matches;

wherein in the event there are no category matches corresponding to the items identified, a request is generated by the system processor for further user input via the speech recognition engine;

wherein in the event there are category matches corresponding to the items identified, and a single category has been returned as a potential match, a request for further sub category and item attributes is made via the speech recognition engine; and

wherein in the event a plurality of categories are returned with more than one corresponding item, a further request is made for a confirmation of a suitable category from the plurality of categories returned via the speech recognition module engine, and the process continues until an item is correctly identified by the system processor.