Construction projects involve multiple company departments and disciplines. The departments follow certain rules in implementing a project, also referred to as requirements in a construction contract. Current administration practices do not show which discipline or department is related to any requirement in the contracts. Thus, all departments need to review contract requirements but typically only from their perspective and with minimal communication with one another. In addition to the tendency of this manual process to error, time and money are lost in evaluating irrelevant departmental requirements. This study concentrates on one aspect of contract interpretation, coordination of the contract requirement review. Automating a classification of the contract requirements by relevant departments can increase the efficiency of contract reviews. This study proposes a robust approach to automating contract sentence classification by relevance to the company department. The approach comprises both natural language processing (NLP) and supervised machine learning techniques to train an algorithm. Training data are selected from an internationally and widely used standard form of construction contract. Precision metric results as high as 0.952 and recall metric results as high as 0.786 are acquired by support vector classifiers (SVCs). These are considered sufficient within the context of multilabel classification of construction contract sentences for construction professionals to operate without further training. The developed methodology reduces time spent on contract review, reliably and accurately predicts classification of contract sentences for departmental relevance, and also removes the dependence on expert participation in coordination efforts contract review.