Beyond tickets and texts : addressing challenges in AI-driven customer service

Abstract

Adapting AI-driven solutions in customer services (CS) is becoming increasingly important to counter the combined impact of increased workload and management expectations. Addressing copious text volumes, text analytics (TA) assumes an important supporting role within CSs. However, the absence of TA and NLP application guidelines slows progress. Utilization of AI advancements in linguistically intricate languages and less digitally developed regions presents challenges. Using a systematic in-depth review of the literature and expert feedback, this study presents the TA/NLP landscape in the context of CS. The paper discusses challenges, open issues, and active research areas in TA/NLP, mainly in customer service, and validates its findings with a country example and industry experts. In addition to the theoretical and practical value of the AI-driven customer service landscape, a unique methodological approach, which includes AI-assisted content mining with BERTopic, systematic literature review, domain expert validation, and the example contributes to methodology development.