Developing large-scale industrial systems requires high-quality requirements to avoid costly rework and project delays. However, linguistic ambiguities in natural language (NL) requirements have been a long-standing challenge, often introducing misinterpretations and inconsistencies that propagate throughout the development lifecycle. Such ambiguous NL requirements necessitate early detection and well-reasoned explanations to clarify and prevent further misunderstandings among stakeholders. While solutions have been developed to detect ambiguities in NL requirements, the advent of generative large language models (LLMs) offers new avenues for explanation-augmented requirements ambiguity detection. This paper empirically investigates LLMs for ambiguity detection and explanation in real-world industrial requirements by adopting an in-context learning paradigm. Our results from three industrial datasets show that LLMs achieve a 20.2% average performance increase in classifying ambiguous requirements when prompted with ten relevant in-context demonstrations (10 -shot), compared to no demonstrations (0 -shot). Additionally, we conducted human evaluations of the LLM-generated outputs with eight industry experts along four dimensions-naturalness, adequacy, usefulness and relevance-to gain practical insights. The results show an average rating of 3.84 out of 5 across evaluation criteria, indicating that the approach is effective in providing supporting explanations for requirement ambiguities.

Allocation of requirements to different teams is a typical preliminary task in large-scale system development projects. This critical activity is often performed manually and can benefit from automated requirements classification techniques. To date, limited evidence is available about the effectiveness of existing machine learning (ML) approaches for requirements classification in industrial cases. This paper aims to fill this gap by evaluating state-of-the-art language models and ML algorithms for classification in the railway industry. Since the interpretation of the results of ML systems is particularly relevant in the studied context, we also provide an information augmentation approach to complement the output of the ML-based classification. Our results show that the BERT uncased language model with the softmax classifier can allocate the requirements to different teams with a 76% F1 score when considering requirements allocation to the most frequent teams. Information augmentation provides potentially useful indications in 76% of the cases. The results confirm that currently available techniques can be applied to real-world cases, thus enabling the first step for technology transfer of automated requirements classification. The study can be useful to practitioners operating in requirements-centered contexts such as railways, where accurate requirements classification becomes crucial for better allocation of requirements to various teams.