Overview

In online customer service, voice recordings provide valuable insights into interactions between employees and customers. Speech analytics - integrating automatic speech recognition, natural language processing (NLP), and acoustic analysis - facilitates the analysis of voice recordings.

Research indicates that in the faceless environment of online customer service, the exchange of genuine emotions acts as the "social glue" that humanizes the interface, transforming a pure information exchange into a meaningful experience that has strong influence on a customer's perception of the service.  Leveraging speech analytics, it is possible to analyze emotions and their impact on employee-customer interactions. In the thesis project, the following research question is addressed: 

How should a speech analytics system for the analysis of emotions in employee–customer interactions be designed?

Goal of the Thesis: 

The thesis focuses on the latest methodologies for investigating customer-employee interactions through speech analytics. The findings will help evaluate where an AI-based prototype can be integrated into customer service. The student will:

• Analyze state-of-the-art speech analytics frameworks in customer service
• Collect requirements and conceptualize a speech analytics system
• Implement the speech analytics system and perform an evaluation

The thesis is done in cooperation with Allianz.

Recommended skills:

• Strong programming skills and experience with data analysis software (i.e., Python, R)
• Familiarity with data science as natural language processing and Generative AI (e.g., clustering, segmentation)
• Strong time management and communication skills
• Strong analytical and English skills

References

Madanian, S.; Parry, D.; Adeleye, O.; Poellabauer, C.; Mirza, F.; Mathew, S.; and Schneider, S., "Automatic Speech Emotion Recognition Using Machine Learning: Digital Transformation of Mental Health" (2022). PACIS 2022 Proceedings. 45. https://aisel.aisnet.org/pacis2022/45 •     Yurtay, Y.; Demirci, H.; Tiryaki, H.; Altun, T. Emotion Recognition on Call Center Voice Data. Appl. Sci. 2024, 14, 9458. https:// doi.org/10.3390/app14209458