Automatic determination of customer sentiment from communications using contextual factors

A computer program product comprising a non-transitory computer-readable medium having computer-readable program code embodied therein to be executed by one or more processors, the program code including instructions to:

• receive communication content exchanged between a customer device and an agent device, the communication content comprising inbound communication to the agent device;
• filter the received communication content to remove predetermined technical support syntax;
• assign baseline values to features within the filtered communication content, the assigning comprising:
• identifying the features from the filtered communication content, the features pertaining to expressed sentiments in the filtered communication content;
• applying a first machine-learning model, which is trained for entity detection to classify each identified feature as relating to one of a product support experience and a product experience;
• applying the first machine-learning model, which is trained to identify polarities and strengths related to identified features, to each product support experience identified feature;
• and
• assigning, by the processor, a score value to each identified product support experience feature, the score values being based on the polarities and strengths identified for each product support experience identified feature;
• weight a subset of the score values based on metadata associated with the filtered communication content;
• and
• combine the score values using a second machine-learning model, which is trained to use weighted combinations to determine an overall sentiment of the filtered communication content.