The key objectives are to identify and evaluate open-source AI tools and models, including large language models (LLMs) and multimodal LLMs, to determine their suitability for supporting deliberation activities, with the potential for fine-tuning existing models or training new ones if necessary. A reusable and modular Reference Architecture for AI-driven deliberations will be designed to ensure scalability and adaptability. The development of a Deliberation Knowledge Graph will further enhance the structuring and integration of deliberative processes. Additionally, it aims to enhance an open-source deliberation platform by incorporating multimodal AI features and gamification elements while also developing a mobile application to broaden accessibility and engagement.