In long-term care (LTC), interviews with clients, their families, and care professionals are essential for understanding and improving care quality. Manual analysis of these interviews is labor-intensive and time-consuming. This study presents an AI-approach, using natural language processing (NLP), to automate: 1) thematic analysis, 2) sentiment detection and 3) summary generation. Using 470 interviews (125 manually labelled by experts as reference) we applied discourse-aware segmentation and contrastive machine learning for context-rich textual understanding. Structured prompting ensured coherent, accurate summaries. The NLP models substantially improved the efficiency and consistency of identifying key themes and sentiment from the interviews. Expert evaluations rated the NLP summaries as coherent and factual. Quantitative evaluations, comparing manual expert analyses to NLP, will be presented at the time of the conference. As an open-source tool, this approach facilitates evidence-informed decision-making in daily practice. Moreover, this model is easily transferable to other types of qualitative data, offering broad applicability across healthcare and research contexts