Enhanced Workflow Management using an Artificial Intelligence ChatBot

Abstract

This paper presents a novel approach to the development of an intelligent workflow engine utilising an artificial intelligence (AI) chatbot. The proposed system combines the advantages of traditional workflow management systems with the power of AI to provide an efficient, user-friendly and highly customisable solution for workflow generation. The chatbot acts as a natural language interface for users to interact with the system, reducing the learning curve and increasing usability. The intelligent workflow engine uses Natural Language Processing (NLP) techniques by applying Large Language Models (LLMs) to enable seamless communication between users and the system. The AI chatbot actively helps users to create workflows based on their needs and supports them in making decisions, streamlining and automatically generating the entire workflow. The new methodology is integrated into an existing software framework for process automation and collaboration in the field of aircraft propulsion systems, extending an existing traditional workflow management system. Full insight into the implementation is provided. To validate the performance and flexibility of the extended workflow engine, the methodology is applied to generic workflows, each representing a unique challenge in workflow architecture. This involves investigating different LLMs and LLM parameters. Finally, the application and evaluation of the methodology is demonstrated by means of a specific application from the field of propulsion technology.