This study aimed to explore the impact of generative artificial intelligence tools on visual identity design through an actual applied experiment and a specialized field evaluation. The study adopted the quasi-experimental, descriptive, and comparative approach. Seven AI platforms were tested, and three complete Arabic visual identity projects were developed: Midad for Digital Education, Wajha for Tourism, and Sa‘ah for the Contemporary Saudi Café. These projects were evaluated by 124 designers and specialists through a questionnaire comprising 12 evaluation criteria, and the data were analyzed using SPSS. The results revealed a consistent and clear pattern across the three projects: AI outputs excelled in execution efficiency—such as color consistency, visual output, and ease of application—with means ranging between 3.43 and 4.08, while showing a noticeable weakness in originality and competitive distinctiveness, with means dropping to 2.75. The overall mean for the three projects was 3.45, rated as “Very Good”. The study also identified a structural shortcoming in handling geometric Arabic calligraphy. One-way ANOVA showed that years of experience do not affect evaluative judgments (Sig. = 0.410), which enhances the reliability of the results and confirms their homogeneity. The study concluded that generative AI is an excellent execution facilitator but not an independent innovator, and that its optimal use is as a supporting tool in the brainstorming stage rather than a final product delivered without creative human review.