With the advancement of generative artificial intelligence, design is moving towards a new era of intelligence. Generative AI models have demonstrated remarkable capabilities in creating images using text prompts, significantly advancing graphic advertisement design and production. Graphic advertisement design has relatively specific business objectives during production, requiring accurate conveyance of the image’s information to users. However, generative AI still faces challenges in accurately generating images that match the textual content, often failing to produce the images designers need. We developed AdvisorAI (Generative AI-Assisted Design System for Graphic Advertising), an intelligent interactive tool designed to assist novice graphic designers in creating graphic advertisements. AdvisorAI features an explanatory interactive design framework, helping users intuitively perceive and modify the generation process. We trained an AI Agent model to analyze users’ natural language inputs and convert them into design feature texts aligned with professional graphic image knowledge. Furthermore, we developed a standard parameter asset library specific to graphic advertisement design, including prompt words, reference images, and generation parameters. These are visualized to aid users in understanding and adjusting the parameters used for image generation, transforming design features into precise parameter combinations for AI-driven image creation. We conducted user studies comparing the workflow using AdisorAI with that of existing industry standard tools, followed by in-depth user interviews. Both quantitative and qualitative research findings indicate that AdvisorAI is highly effective in assisting novice graphic designers to quickly utilize generative AI for graphic design production.