ORGANIZATIONAL INTELLIGENCE EXTRACTION FROM MEETING TRANSCRIPTS

Abstract: Meetings generate large volumes of conversational data, but critical information such as action items, responsibilities, and decisions often gets buried in lengthy discussions. This paper proposes a Natural Language Processing (NLP)-based system called Organizational Intelligence Extraction from Meeting Transcripts that automatically extracts actionable insights from meeting transcripts. The system accepts both text and audio inputs, preprocesses the transcript text using NLTK, classifies sentences into action or decision categories, applies Named Entity Recognition (NER) using spaCy to identify responsible persons and deadlines, and generates structured tabular output. Unlike traditional meeting summarization tools, this approach focuses specifically on structured extraction of tasks and decisions to support project tracking and team coordination. The system is implemented as a Flask web application and evaluated on sample meeting transcripts, demonstrating its capability to accurately identify and organize actionable information.

Keywords: Natural Language Processing, Named Entity Recognition, Meeting Transcript Analysis, Action Item Extraction, Decision Detection, Text Classification, NLTK, spacy, Flask.