Co-Pilot for Project Managers: Developing a PDF-Driven AI Chatbot for Facilitating Project Management

摘要

Our AI-driven PDF Chatbot is specialized for Project Management (PM) Automation and acts as a virtual Project Manager that offers continuous support to global teams. It interprets PDF data like SRS reports and interview transcripts by utilizing Open-Assistant’s SFT-1 12B Model. Insights from interviews of 15 project managers have enriched the knowledge base of our chatbot and ultimately enabled informative responses to the stakeholders of the project. Advanced AI techniques ensure efficient text preprocessing, including tokenization, numerical normalization, lowercasing, removing punctuation, removing extra spaces, recursive character text splitter, and lemmatization. It is primarily tailored for e-commerce project and provides precise guidance based on e-commerce data and risk management factors. With an average cosine similarity of 80.80% and semantic similarity score of 85.21%, it consistently aligns with PDF Contents and optimize the project management phases & methodologies. This innovation enhances Human-Robot Interaction, PM Automation, facilitates decision-making, and enables uninterrupted communication. While AI-driven PDF chatbots like ChatPDF and SciSummary exist, our chatbot is uniquely focused on automating project management tasks, providing tailored insights for e-commerce projects and decision-making, thus offering a breakthrough in PM automation. To ensure the chatbot’s robustness in context-aware responds, we compare our chatbot with ChatPDF and Sci-summary which are some PDF driven chatbots. Making our work available open-source on <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/codewithkhurshed/SPM-project-repo</uri> can enhance its accessibility and promote future research opportunities in PDF driven chatbot development.