In large-scale projects, the client defines a set of product requirements, which the provider is then expected to deliver within the agreed time, cost and scope. If a client sets ambiguous requirements for the project, this may result in the receipt of an unsatisfactory product. Therefore, reducing ambiguity in product requirements by the client is a critical success factor. Despite its significance and regular occurrence, the requirement ambiguity problem has not yet received a methodological solution that can fit large-scale projects, which commonly include a great number of requirements. This paper proposes a semi-automated approach, which combines natural language processing (NLP) to identify ambiguous terms and statements and a soft computing technique to specify these terms using fuzzy set theory. This work contributes to the current literature on requirement specification by highlighting a line of research which paves the way to leverage the applications of advanced tools to allow the clarification of ambiguous requirements in large-scale projects.