Smart product configurator applying large language models: the case of construction equipment machines

Matching suitable machines to use cases is a recurring challenge for the technical equipment industry. Despite growing interest in large language models (LLMs), their application in manufacturing remains underexplored. This paper presents a smart product configurator that enhances the construction equipment recommendation and sales process, which entails extensive technical knowledge of machine usage and performance. We present the multi-agent configurator architecture, combining retrieval-augmented generation and prompt engineering feeding a LLM to collect data and recommend machine configurations. Initial tests show promising results, highlighting its potential both as a standalone end-user service and as a supporting tool for sales consultants.