Construction site accidents are a recurring concern, resulting in significant human and economic losses. The International Organization (ILO) reports a troubling increase in incident rates and severe outcomes. This underscores the need for improved and more effective incident reporting systems. Traditional incident reporting methods rely on manual data documentation and address issues promptly. This delayed reporting of incidents can cause reduced ability to investigate and learn from the incident, thereby increasing the risk of future incidents and accidents, thus making automated system is essential. This study proposes an advanced automated incident reporting system using Natural Language Processing (NLP). The system utilizes speech recognition technology to record voice reports from workers. NLP algorithms analyze the context and extract relevant entities in real time. A literature review reveals the potential of NLP in improving safety management practices. The review highlights benefits including enhanced accuracy, reduced reporting time and improved documentation. The proposed system involves data collection, speech recognition model training and NLP based incident detail extraction. Automated report templates facilitate rapid incident documentation. The system aims to speed up construction site incident report documentation.