The widespread availability of generative artificial intelligence (AI) is reshaping assessment practices in higher education. Rather than attempting to exclude AI from examinations, this paper argues for assessment designs that deliberately embed AI within human-centered learning processes. Grounded in the concept of deep understanding (Biggs, 1996), the focus lies on students’ capacity to integrate knowledge, exercise judgement, engage in dialogue, and justify context-sensitive decisions. The paper presents a multi-modal online examination implemented in the master’s program in Business Education at WU Vienna.

First introduced in winter semester 2024/25 and redesigned in 2025/26, the case-based assessment combined written analysis, oral and dialogical elements, and reflective tasks. AI use was permitted but required critical review and responsible integration. Across both implementations, formats such as videotaped presentations, peer discussions, task design, and examination dialogues emphasized reasoning, interaction, and reflection. Student feedback and examiner observations were highly positive.

The findings highlight three implications: established principles such as constructive alignment remain essential; deep understanding requires tasks that resist purely algorithmic solutions; and acknowledging generative AI enables transparent, multi-stage assessments that meaningfully capture learning in contemporary higher education.