LLM Support for Real-Time Technical Assistance

Vincenzo Scotti, Mark James Carman

Abstract: In this paper, we present a demo web application that adopts Large Language Models (LLMs) to enhance user support across various fields. Its primary goal is to enable experts, like technicians, to deliver remote assistance more effectively by leveraging LLM capabilities. The application permits experts to browse through a database of documents, including past support chats and manuals, and suggests responses based on previous interactions. We developed the demo using publicly available data sets from technical support and tutoring domains to showcase its adaptability. Key features include search functionality, response suggestions, and automatic information extraction. The demo highlights the potential of LLMs in improving technical support workflows by streamlining knowledge retrieval and aiding technicians in resolving queries, leading to enhanced efficiency and user satisfaction in support interactions.

External IDs:dblp:conf/pkdd/ScottiC24