Students are entering a profession where they will work alongside AI systems, so architectural education cannot ignore these tools and must change how it teaches. Design studios have long rewarded production over testing. Because modeling, analysis, and rendering take so long, students tend to commit early to the first workable idea and spend the rest of the term producing it. Facades, interiors, and building performance are often decided by default. Herbert Simon called this satisficing: stopping at “good enough” when search is costly. This paper argues that generative AI reduces production effort, so testing can begin earlier and continue longer. Drawing on a new seminar and a parallel design studio using Autodesk Forma, TestFit, and Nano Banana (Google Gemini), it presents project cases. On a steep site, one student set aside TestFit’s answer when it fell short of her design intent, while another reframed the site so the tool could help. In a housing studio, students given the same feedback took different paths: one team declined AI to keep their authorship, while a seminar student layered AI on his own work, caught a convincing but wrong AI-generated section, and decided what to do himself and what to hand off. A thesis student went further, putting his design intent into what he gave the AI by drawing it, not just describing it. Cheaper testing is not automatically better design. What let students judge the tools was what they already knew about architecture. The paper proposes a teaching principle: agency is the deliverable, meaning a decision the student can stand behind, even when that decision is to set the technology aside. Four simple questions make that judgment visible in any discipline: Why this option? What did you compare? What did you give up? What would change your mind?
Murali Paranandi is a Professor of Architecture at Miami University, where he has taught for nearly three decades. He holds a Master of Architecture from The Ohio State University, where he trained under Dr. Chris Yessios, creator of formZ. His research focuses on engaging emerging digital media in design, with recent work on how generative AI tools—including Forma, TestFit, and Gemini—are reshaping architectural pedagogy. His students’ work has won design awards and been exhibited internationally. He was recognized by Autodesk as an Exemplar Educator.