Assessments play a crucial role in student learning, yet designing assessments that accurately evaluate deep learning and higher-order thinking skills remains challenging for educators. Traditional methods often fail to capture authentic understanding, favouring surface-level memorization. This issue is further compounded by Artificial Intelligence (AI) language models that can potentially assist students. This work-in-progress paper reports on the first phase of a more extensive study conducted at a Faculty of Engineering (FoE) at a Higher Education Institution (HEI). The study is aimed at developing a toolkit to support educators in assessment design for the digital age. In this phase semi-structured interviews were performed with three engineering educators to explore their practices and broad perceptions around assessing deep learning. Findings from these interviews will inform the design of subsequent phases which include a survey for all faculty staff, focus groups with educators and consultations with educational and AI experts. The expected outcome is of the full study is a versatile toolkit enabling educators to critically evaluate assessments, align them with fostering deep learning, and adapt to AIassisted education. The toolkit will provide a framework for analysing assessments, identifying shortcomings, and incorporating strategies to promote authentic learning experiences while mitigating AI influence. This research contributes to effective assessment practices and responsible AI integration in education.