Information provision for call centres

• US 20050216269A1
• Filed: 07/25/2003
• Published: 09/29/2005
• Est. Priority Date: 07/29/2002
• Status: Active Grant

• Current Assignee: British Telecommunications PLC (BT Group PLC)
• Original Assignee: British Telecommunications PLC (BT Group PLC)
• Inventors: Ringland, Simon Pa, Deans, Paul B., Scahill, Francis J.
• Granted Patent: US 7,542,902 B2
• Time in Patent Office: Days
• US Class Current: 704/270.100
• CPC Class Codes: G06F 16/3334 Selection or weighting of terms from queries, including natural language queries
• CPC Class Codes: G10L 15/1815 Semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning
• CPC Class Codes: H04M 2201/40 using speech recognition speech recognition per se G10L15/00
• CPC Class Codes: H04M 2203/357 Autocues for dialog assistance
• CPC Class Codes: H04M 3/5183 Call or contact centers with computer-telephony arrangements

Abstract

The invention provides for a voice platform to monitor a conversation between a call centre agent and a caller to identify any predetermined keywords or phrases used in the conversation therebetween. These keywords or phrases can then be used to interface into an existing knowledge management system in order to allow information from the system to be pushed to the agent, thus improving agent efficiency. In a preferred embodiment additional processing is provided which generates an information item score based in the detected keywords or phrases, and those information items with the highest scores are then pushed to the agent, by displaying shortcuts to the information to the agent in the form of a shortcut tree.

Claims

1. An information interface system, comprising:

speech recognition means arranged to listen to voice signals carried on a communications channel between a user and another person, and to recognise at least one or more predefined keywords or phrases contained within the voice signals;

keyword processing means arranged to relate any recognised keywords or phrases to items of information stored in an information database so as to identify relevant items of information thereto;

information display means arranged to display any items of information and/or links thereto identified by the keyword processing means to the user; and

speech recognition control means arranged to activate or deactivate the speech recognition means in response to one or more predetermined criteria.

2. A system according to claim 1, wherein the keyword processing means further comprises:

storage means for storing item definition data defining the items of information in the information database;

item matching means for matching recognised keywords or phrases to the item definition data; and

item scoring means for keeping an item score for each item of information in dependence upon the number of keywords or phrases matched to each item.

3. A system according to claim 2, wherein the keyword processing means further comprises item sorting means arranged to sort the items of information in dependence on the respective item scores.

4. A system according to claim 2, wherein the keyword processing means further comprises item score thresholding means arranged to apply at least one threshold to the item scores, wherein those items whose item scores do not meet the threshold are not made available to the user.

5. A system according to claim 1, wherein the information display means is arranged to display the items of information and/or links thereto as a hierarchical structure, which preferably matches the hierarchy of the information database.

6. A system according to claim 1 wherein the speech recognition means, the keyword processing means and the information display means are substantially continuously or periodically operable so as to continually or periodically update the items of information and/or links thereto displayed to the user.

7. A system according to claim 1, wherein the predetermined criteria are selected such that the speech recognition means is deactivated on that portion or portions of the voice signals which are not expected to contain keywords or phrases.

8. A method of interfacing information to a user, comprising the steps of:

listening to voice signals carried on a communications channel between the user and another person;

recognising, in response to one or more predetermined criteria, at least one or more predefined keywords or phrases contained within the voice signals;

relating any recognised keywords or phrases to items of information stored in an information database so as to identify relevant items of information thereto; and

displaying any items of information and/or links thereto so identified to the user.

9. A method according to claim 8, wherein the relating step further comprises:

storing item definition data defining the items of information in the information database;

matching recognised keywords or phrases to the item definition data; and

keeping an item score for each item of information in dependence upon the number of keywords or phrases matched to each item.

10. A method according to claim 9, wherein the relating step further comprises sorting the items of information in dependence on the respective item scores.

11. A method according to claim 8, wherein the relating step further comprises applying at least one threshold to the item scores, wherein those items whose item scores do not meet the threshold are not made available to the user.

12. A method according to claim 8, wherein the displaying step is arranged to display the items of information and/or links thereto as a hierarchical structure, which preferably matches the hierarchy of the information database.

13. A method according to claim 8, wherein the steps are substantially continuously or periodically repeated so as to continually or periodically update the items of information and/or links thereto displayed to the user.

14. A method according to claim 8, wherein the predetermined criteria are selected such that the recognising step is not performed on that portion or portions of the voice signals which are not expected to contain keywords or phrases.