By Jing Qing, Veronica Bloom, Michael Addonisio, Christy Gearheart; Charter Communications
The goal of this project is to utilize artificial intelligence (AI) and machine learning (ML) techniques to gain insights into the drivers of support calls. These insights are used to determine areas of improvements to both processes and products to enhance our customer experience.
In this paper, we discuss the end-to-end automated pipeline to go from call center data to actionable insights utilizing machine learning techniques such as Automatic Speech Recognition (ASR) and Natural Language Processing (NLP). Discussion of the pipeline architecture, use cases of automated call disposition, and call topic trend analysis are also included. Broadly, the pipeline converts recorded call audio conversations to text transcripts using automatic speech recognition, then uses those transcripts to train classification models to predict call disposition. Unsupervised topic modeling extracts new trends and topics from the call-volume data over time to identify new or emerging call drivers. These call drivers subsequently drive new feature development and product roadmaps.