Volunteering holds a vital place in enhancing the quality of life for individuals in need. Online volunteering platforms have emerged as valuable tools for connecting volunteers with causes that require assistance. The local civic engagement organization used in this study is one of South Africa’s leading online volunteering platforms, which enables causes to list their needs and volunteers to fulfill them. This study proposed using machine learning for sentiment analysis to analyze messages sent to causes by volunteers expressing their commitment to a particular need. Long Short-Term Memory (LSTM) was used in this study to analyze these messages in conjunction with their associated ratings or sentiments. The findings of this study demonstrate that Long Short-Term Memory (LSTM) exhibits superior performance in predicting online volunteering commitment compared with Gated Recurrent Unit (GRU) and Simple Recurrent Neural Network (SRNN). These findings directly affect the local civic engagement organisation as an online volunteering platform because it leverages these findings to enhance its volunteer engagement strategies. Specifically, the local civic engagement organisation can utilise these insights to identify volunteers with a high likelihood of not committing and tailor its communication approaches accordingly. This targeted approach can foster greater satisfaction among volunteers and causes, ultimately driving increased traffic to the Local civic engagement organisation platform and bolstering its reputation as a premier destination for online volunteering.