AI-Driven Real-Time Summarization and Action Item Extraction in Video Conferencing Platforms

Abstract

The sudden transition to digital cooperation greatly contributed to the high use of video conferencing systems in the sphere of business, education, and healthcare. Nevertheless, the sheer amount of data in the form of conversations that are produced during meetings is frequently inefficient, and participants find it difficult to summarize the most important information and implement the decisions. The current paper introduces an AI-based platform of real-time summarization and action item extraction within the video conferencing setting. The system proposed is an automatic speech recognition system that incorporates high-level natural language processing algorithms to create summaries of the meetings and identify and organize actionable items. The framework offers solutions to the issues of speech disfluency, multi-speaker conversations, and domain variation by relying on transformer-based architectures and context-aware learning. The experimental outcomes on benchmark meeting datasets show an increase in the accuracy of the summarization and task identification of experimental results over the baseline solution. Moreover, the combination of real-time summarization with a conferencing platform reveals the possibility of improved productivity, accountability, and knowledge retention. The study is relevant to the emerging body of AI-based collaboration tools and it provides a solution that is scalable and practical to enhance productivity in remote and hybrid work environments.