Techniques are presented herein that support the automated reorganization of a meeting transcript. According to the presented techniques, when a meeting is recorded it is first transcribed through an automated speech-to-text system. Then, the resulting raw document is decomposed into sections corresponding to different topics, the topics are reorganized into a more coherent and intelligible structure, and missing pieces of information may be identified and then added to the document. Such postprocessing increases the intelligibility and value of meeting transcriptions and ensures that they remain understandable and useful over the long term. Aspects of the presented techniques may leverage large language models (LLMs) and automatic speech recognition (ASR) capabilities.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.