Addressing complex global challenges, such as sustainability and the circular economy, demands interdisciplinary collaboration. However, such efforts are often hindered by knowledge silos, value conflicts, and communication barriers. This paper presents a novel co-design platform that integrates a human designer with a team of multi-agents powered by Large Language Models (LLMs). Grounded in the Double Diamond model, the platform facilitates iterative cycles of divergence and convergence, leveraging design cards to encapsulate and structure interdisciplinary insights into accessible formats. The multi-agent system includes LLM-based agents specializing in distinct domains—environment, technology, and psychosocial factors—enabling targeted and holistic problem-solving. A case study exploring circular economy solutions demonstrates the platform’s efficacy in bridging knowledge gaps, resolving value conflicts, and enhancing creativity. Results show that human-guided AI collaboration fosters a dynamic and scalable co-design process, offering actionable solutions to complex interdisciplinary challenges.