In recent years, the rapid advancement of generative AI tools—such as DeepSeek, ChatGPT, and ChatPPT—has profoundly transformed university teaching models. Leveraging its powerful capabilities in deep learning and natural language processing, AI empowers classroom instruction across multiple dimensions—including content generation, personalized learning, intelligent tutoring, data analysis, and virtual teaching—providing new perspectives and practical pathways for teaching undergraduate finance and taxation courses. Taking the core finance and taxation course “Tax Law” as a case study, this paper proposes utilizing AI tools to implement pre-class diagnostic assessments of student learning status and resource updates; facilitate in-class scenario creation and intelligent interaction; and enable post-class dynamic evaluation and consolidation/extension of knowledge. This approach aims to shift teaching from being experience-driven to data-driven, and to transform teaching evaluation from focusing solely on final exam results to encompassing a diverse range of learning processes, thereby offering a practical pathway for addressing the fragmentation issue between “teaching, learning, and assessment” in finance and taxation courses.