Public procurement tender documentation may contain a large amount of information distributed across several different documents. Finding and correctly understanding this information can therefore be time-consuming. The aim of this thesis is to investigate the usefulness of large language models in the analysis of public procurement tender documentation. The thesis presents the basic capabilities and limitations of artificial intelligence, large language models, and retrieval-augmented generation (RAG). In addition, the main characteristics of public procurement and tender documentation are presented. The central part of the thesis focuses on the analysis of the effectiveness of large language models when working with tender documentation. The study was conducted with eight participants divided into two groups. One group answered the questions using a large language model, while the other group answered them without its use. The groups were compared in terms of completion time, accuracy of answers, and participants’ observations. On average, participants using a large language model completed the questionnaire faster and achieved higher answer accuracy. Their observations also indicate that finding information was easier. The results show that large language models can be a useful support tool when working with tender documentation. Nevertheless, it is important that users verify the obtained information against the original documentation and retain final judgement.
|
