Identifying actionable insights in unstructured datatypes of a semantic knowledge database

1 . A system for analyzing customer feedback data received via multiple input channels, said system comprising:

a computer with one or more processors and memory, where the computer is configured to process and analyze customer feedback data received from a plurality of external data sources via the multiple input channels; and

a network connection operatively connecting the external data sources to the computer,

where the computer is configured to perform steps including:

importing the customer feedback data from the external data sources into a data library, where the multiple input channels include voice of the customer complaints, online chat transcripts, phone call transcripts, social media posts, ratings and comments from mobile application stores and customer feedback provided via a third-party business;

pre-processing the customer feedback data in the data library to provide a cleaned dataset, where the pre-processing includes removing personally identifiable information, simplifying language, consolidating terms, removing stop words and removing noise from the customer feedback data;

loading the cleaned dataset into a semantic knowledge database, including transforming the cleaned dataset into a standardized data structure;

analyzing the semantic knowledge database, using a machine learning algorithm including a natural language processing (NLP) application and a segmentation application, to produce output data including a use case database; and

providing the use case database to a set of consumption applications, where the set of consumption applications are utilized by users to identify actionable insights, where the actionable insights include customer sentiments, and clusters of products, services and business segments appearing in the customer feedback data and associated with the customer sentiments;

wherein the NLP application is trained via supervised learning using a training dataset of customer feedback records labeled with corresponding actionable insights, wherein the supervised learning causes neural networks in the NLP application to establish layers, nodes, connectivity, and weighting, which minimize error signals between the labeled customer feedback records and the corresponding actionable insights in the training dataset;

wherein the semantic knowledge database is updated by loading a new version of the cleaned dataset, and the updated semantic knowledge database is analyzed by the machine learning algorithm including the NLP application to produce new output data, when an elapsed time from a previous update of the semantic knowledge database exceeds a threshold or a data update triggering event occurs; and

wherein the NLP application receives updated training using a new training dataset resulting in an updated NLP application, and the updated NLP application in the machine learning algorithm is used to analyze a latest version of the semantic knowledge database to produce new output data, the updated training of the NLP application occurring when an elapsed time from a previous training of the NLP application exceeds a threshold or an NLP update triggering event occurs.

2 . The system according to claim 1 wherein, after identifying actionable insights, the set of consumption applications are utilized by the users to trace the customer sentiments and the clusters of products, services and business segments back to original records in the customer feedback data received from the plurality of external data sources via the multiple input channels.

3 . A method for analyzing customer feedback data received via multiple input channels, said method comprising:

importing the customer feedback data from a plurality of external data sources into a data library, where the multiple input channels include voice of the customer complaints, online chat transcripts, phone call transcripts, social media posts, ratings and comments from mobile application stores and customer feedback provided via a third-party business;

pre-processing the customer feedback data in the data library, using a computer having a processor and memory, to provide a cleaned dataset, where the pre-processing includes removing personally identifiable information, simplifying language, consolidating terms, removing stop words and removing noise from the customer feedback data;

loading the cleaned dataset into a semantic knowledge database, including transforming the cleaned dataset into a standardized data structure;

analyzing the semantic knowledge database, using a machine learning algorithm running on the computer, said machine learning algorithm including a natural language processing (NLP) application and a segmentation application, to produce output data; and

providing the output data to a set of consumption applications, where the set of consumption applications are utilized by users to identify actionable insights, where the actionable insights include customer sentiments, and clusters of products, services and business segments appearing in the customer feedback data; and

wherein the NLP application receives updated training using a new training dataset resulting in an updated NLP application, and the updated NLP application in the machine learning algorithm is used to analyze a latest version of the semantic knowledge database to produce new output data, the updated training of the NLP application occurring when an elapsed time from a previous updated training of the NLP application exceeds a threshold or an NLP update triggering event occurs.

4 . The method according to claim 3 wherein the multiple input channels include voice of the customer complaints, online chat transcripts, phone call transcripts, social media posts, ratings and comments from mobile application stores and customer feedback provided via third-party businesses.

5 . The method according to claim 4 wherein, after identifying actionable insights, the set of consumption applications are utilized by the users to trace the customer sentiments and the clusters of products, services and business segments back to original records in the customer feedback data received from the plurality of external data sources via the multiple input channels.

6 . The method according to claim 4 wherein the phone call transcripts are derived from audio data collected during customer phone calls, where a speech signal processing system performs audio transduction on the audio data to generate the phone call transcripts in the form of digital text data.

7 . The method according to claim 3 wherein the semantic knowledge database is updated by loading a new version of the cleaned dataset, and the updated semantic knowledge database is analyzed by the machine learning algorithm including the NLP application to produce new output data, when an elapsed time from a previous update of the semantic knowledge database exceeds a threshold or a data update triggering event occurs.

8 . A method for analyzing customer feedback data received via multiple input channels, said method comprising:

importing the customer feedback data from a plurality of external data sources into a data library, where the multiple input channels include voice of the customer complaints, online chat transcripts, phone call transcripts, social media posts, ratings and comments from mobile application stores and customer feedback provided via a third-party business;

pre-processing the customer feedback data in the data library, using a computer having a processor and memory, to provide a cleaned dataset, where the pre-processing includes removing personally identifiable information, simplifying language, consolidating terms, removing stop words and removing noise from the customer feedback data;

loading the cleaned dataset into a semantic knowledge database, including transforming the cleaned dataset into a standardized data structure;

analyzing the semantic knowledge database, using a machine learning algorithm running on the computer, said machine learning algorithm including a natural language processing (NLP) application and a segmentation application, to produce output data; and

providing the output data to a set of consumption applications, where the set of consumption applications are utilized by users to identify actionable insights, where the actionable insights include customer sentiments, and clusters of products, services and business segments appearing in the customer feedback data; and

wherein the semantic knowledge database is updated by loading a new version of the cleaned dataset, and the updated semantic knowledge database is analyzed by the machine learning algorithm including the NLP application to produce new output data, the semantic knowledge database being updated when an elapsed time from a previous update of the semantic knowledge database exceeds a threshold or a data update triggering event occurs.

9 . The method according to claim 8 , wherein the multiple input channels include voice of the customer complaints, online chat transcripts, phone call transcripts, social media posts, ratings and comments from mobile application stores and customer feedback provided via third-party businesses.

10 . The method according to claim 8 , wherein, after identifying actionable insights, the set of consumption applications are utilized by the users to trace the customer sentiments and the clusters of products, services and business segments back to original records in the customer feedback data received from the plurality of external data sources via the multiple input channels.

11 . The method according to claim 8 , wherein the phone call transcripts are derived from audio data collected during customer phone calls, where a speech signal processing system performs audio transduction on the audio data to generate the phone call transcripts in the form of digital text data.