<p>Generative Artificial Intelligence (GenAI) introduced new possibilities and challenges into university classrooms across various disciplines. In particular, the text-image generation technology has significantly impacted architecture and design education throughout the whole process, from initial idea brainstorming to the manufacturing of the final artifact products. Based on a mixed-method case study of a Project-based Learning (PBL) Design Drawing course for first-year students who majored in architecture in a Chinese university, this paper examines how a text-image GenAI tool, <i>Midjourney,</i> is integrated into the course to improve student learning experiences and outcomes. The key findings suggest that the integration of the GenAI tool in the course has a positive impact on both students’ learning experiences and outcomes. Notably, students with limited foundational competencies benefited more significantly from using the tool; it may have bridged the competency gaps among peers. Additionally, a positive correlation is observed between the frequency of GenAI usage and improved learning outcomes. <i>Midjourney</i> enhanced students’ artistic competencies, aesthetic sensibility, and confidence. Identified challenges from the case study encompassed equitable access to GenAI technologies, the extent of teacher intervention, and <i>Midjourney's</i> technical drawbacks. This exploration offers a practical example for advancing pedagogical approaches and fostering creativity and innovation in architecture education and relevant fields. It contributes to evolving educational paradigms and provides actionable pathways for reconstructing design pedagogy across related disciplines.</p>