Granted Patent B2

US 12,731,156 · App. 17/066,967 · Granted Sep 8, 2026

Inventor: Scott Mackie (Glasgow, GB)

Assignee: Verint Americas Inc.

G06Q30/016G06F16/93G06Q30/02H04M3/5175

Abstract

Claims (51)

1 . A method for machine learning to generate analytics rules for predicting one or more computerized resource modules for use in completing an interaction, comprising:

receiving a plurality of historic interactions and interaction metadata as interaction data structures, each interaction data structure including a plurality of fields, including an interaction data field, an internal contact data field, customer identification data fields, and a representative electronic resource data field listing computerized resource modules used by a customer service representative associated with the internal contact data field for a historic interaction;

performing agent desktop use analytics and text analytics or voice analytics on each interaction data structure, wherein the agent desktop use analytics include monitoring use of agent desktops, and receiving a set of data pertaining to use of agent desktops associated with each interaction data structure of the interaction data structures;

generating a customer sentiment based on the agent desktop use analytics, at least one of the plurality of fields in an interaction data structure, and the text analytics or voice analytics;

generating a success indicator for each computerized resource module included in the representative electronic resource data field based on the customer sentiment, the agent desktop use analytics, at least one of the plurality of fields in the interaction data structure, and the text analytics or voice analysis, wherein the success indicator includes a degree of success associated with using a computerized resource module for the corresponding interaction data structure;

generating at least one model configured to create at least one analytics rule using machine learning techniques based on the interaction data structures, generated customer sentiment, and generated success indicator, wherein the at least one analytics rule causes at least one of a plurality of computerized resource modules to be selected to be accessed by a central analytics engine in connection with receipt of a new interaction, further wherein the plurality of computerized resource modules are computerized applications;

using the at least one model to create at least one analytics rule,

wherein the at least one analytics rule is a complex analytics rule corresponding to at least two of the following: a customer tier value, a customer sentiment, a customer feedback score, a type of interaction, a computerized resource module rating, or a set of keywords, and

wherein a category of the complex analytics rule indicates a mobile phone issue for a given customer tier value and a given customer sentiment value, and in response, the complex analytics rule indicates to provide a model article for troubleshooting the mobile phone issue from a resource database of the at least one model to the customer service representative;

monitoring for new interactions data structures and automatically continually updating the at least one model when a new interaction data structure is received based on analysis of interaction data structures such that the at least one model includes the new interaction data structure when creating analytics rules; and

using the updated at least one model to create at least one new analytics rule.

2 . The method of claim 1 , further comprising performing customer value analytics on each interaction data structure and generating a customer value level.

3 . The method of claim 2 , wherein the customer value level is also used in generating the at least one model.

4 . The method of claim 1 , further comprising receiving a new interaction data structure and using the at least one model with the new interaction data structure to predict at least one analytics rule to apply to the new interaction data structure.

5 . The method of claim 1 , wherein the at least one analytics rule is a complex analytics rule corresponding to at least one of the plurality of fields and at least one of the customer sentiment or the success indicator.

6 . The method of claim 1 , wherein the plurality of computerized resource modules include voice analytic applications, text analytic applications, quality assurance applications, and scheduling applications.

7 . A method for completing a real-time interaction using machine learning to predict analytics rules for the real-time interaction, comprising:

receiving a real-time interaction and associated interaction metadata in real-time;

populating a real-time interaction data structure with the real-time interaction and associated interaction metadata, where the real-time interaction data structure includes a plurality of fields, including an interaction data field, an internal contact data field, and customer identification data fields;

performing text analytics or voice analytics on the real-time interaction data structure;

generating a customer sentiment based on at least one of the plurality of fields in the real-time interaction data structure and the text analytics or voice analytics;

inputting the real-time interaction data structure and generated customer sentiment into a trained model to predict at least one analytics rule to be applied to the real-time interaction, wherein:

the at least one analytics rule identifies one or more computerized resource modules of a plurality of computerized resource modules to automatedly execute in real-time while completing the real-time interaction, further wherein the plurality of computerized resource modules are computerized applications,

the at least one analytics rule is a complex analytics rule corresponding to at least two of the following: a customer tier value, a customer sentiment, a customer feedback score, a type of interaction, a computerized resource module rating, or a set of keywords, and

a category of the complex analytics rule indicates a mobile phone issue for a given customer tier value, a given customer sentiment, and a negative customer feedback score, and in response, the complex analytics rule indicates to provide a customer service representative with an electronic resource to identify and fix the mobile phone issue;

applying the at least one analytics rule predicted by the trained model to the real-time interaction data structure and the generated customer sentiment;

executing, by a central analytics engine, in real-time, the one or more computerized resource modules identified by the at least one analytics rule to complete the real-time interaction;

generating a success indicator for each computerized resource module used to complete the real-time interaction, wherein the success indicator includes a degree of success associated with using a computerized resource module for the corresponding real-time interaction data structure;

monitoring for the real-time interaction and automatically continually updating at least one model when the real-time interaction data structure is received based on analysis of the real-time interaction data structure such that the at least one model includes the real-time interaction data structure to create analytics rules; and

using the updated at least one model to create at least one new analytics rule.

8 . The method of claim 7 , further comprising performing customer value analytics on the real-time interaction data structure and generating a customer value level.

9 . The method of claim 8 , wherein the customer value level is also input into the trained model and used by the trained model to predict at least one of a plurality of analytics rules to be applied to the real-time interaction.

10 . The method of claim 7 , further comprising including agent desktop analytics and the one or more computerized resource modules run for the real-time interaction in the fields of the real-time interaction data structure in a representative electronic resource data field.

11 . The method of claim 10 , wherein the success indicator is based on and the customer sentiment, the agent desktop analytics, and fields in the real-time interaction data structure.

12 . The method of claim 7 , further comprising using the updated at least one model on a new real-time interaction to predict at least one analytics rule including the new analytics rule to apply to the new real-time interaction.

13 . The method of claim 7 , further comprising using the updated at least one model to update the at least one analytics rule.

14 . A system for machine learning to generate analytics rules for predicting one or more computerized resource modules for use in completing an interaction, comprising:

a memory comprising computer readable instructions;

a processor configured to read the computer readable instructions that when executed causes the system to:

receive a plurality of historic interactions and interaction metadata as interaction data structures, each interaction data structure including a plurality of fields, including an interaction data field, an internal contact data field, customer identification data fields, and a representative electronic resource data field listing computerized resource modules used by a customer service representative associated with the internal contact data field for a historic interaction;

perform agent desktop use analytics and text analytics or voice analytics on each interaction data structure, wherein the agent desktop use analytics include monitoring use of agent desktops, and receiving a set of data pertaining to use of agent desktops associated with each interaction data structure of the interaction data structures;

generate a customer sentiment based on the agent desktop use analytics, at least one of the plurality of fields in an interaction data structure, and the text analytics or voice analytics;

generate a success indicator for each computerized resource module included in the representative electronic resource data field based on the customer sentiment, the agent desktop use analytics, at least one of the plurality of fields in the interaction data structure, and voice analytics or text analytics, wherein the success indicator includes a degree of success associated with using a computerized resource module for the corresponding interaction data structure;

generate at least one model configured to create at least one analytics rule using machine learning techniques based on the interaction data structures, generated customer sentiment, and generated success indicator, wherein the at least one analytics rule causes at least one of a plurality of computerized resource modules to be selected to be accessed by a central analytics engine in connection with receipt of a new interaction, further wherein the plurality of computerized resource modules are computerized applications;

use the at least one model to create at least one analytics rule,

wherein the at least one analytics rule is a complex analytics rule corresponding to at least two of the following: a customer tier value, a customer sentiment, a customer feedback score, a type of interaction, a computerized resource module rating, or a set of keywords, and

wherein a category of the complex analytics rule indicates a mobile phone issue for a high-tier customer tier value and an unfavorable customer sentiment, and in response, the complex analytics rule indicates to provide an open scheduling electronic resource for a nearest shop for scheduling of an appointment;

monitor the system for new interaction data structures and automatedly continually updating the at least one model when a new interaction data structure is received based on analysis of interaction data structures such that the at least one model includes the new interaction data structure when creating analytics rules; and

use the updated at least one model to create at least one new analytics rule.

15 . The system of claim 14 , wherein the processor is further configured to perform customer value analytics on each interaction data structure and generating a customer value level.

16 . The system of claim 14 , wherein the at least one analytics rule is a complex analytics rule corresponding to at least one of the interaction data field and at least one of the customer sentiment or the success indicator.

Assignments (2)

Continuity (2)

References Cited (14)