Enhancing the industrial design process with generative AI
Fattah Saleh, Helan (2025)
Fattah Saleh, Helan
2025
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2025051311216
https://urn.fi/URN:NBN:fi:amk-2025051311216
Tiivistelmä
This thesis explores the integration of generative Artificial Intelligence (AI) tools into the industrial design process, focusing on how these technologies can support and enhance each phase of the Double Diamond framework: Discover, Define, Develop, and Deliver. As generative AI becomes increasingly accessible and capable, its potential to assist designers with research, ideation, visualization, and modeling continues to grow.
The study investigates how tools such as ChatGPT, Adobe Firefly, and AI-based 3D model generators (like Meshy) can be applied across the design workflow. In the early phases, ChatGPT is used to simulate product benchmarking and create a user persona based on Finnish market expectations. Adobe Firefly is used to generate a visual moodboard aligned with that persona. A survey is also conducted to gather insights from design students and professionals on how AI is perceived in the field. In the practical component, an outdoor light post concept is developed using generative AI tools, while a parallel version is modeled manually in Blender. This enables a comparative analysis in terms of time efficiency, mesh quality, and editability.
By analyzing both research and modeling outcomes, the thesis highlights the strengths and limitations of current AI tools in industrial design. It argues that while AI can meaningfully accelerate and inspire early-stage tasks, manual expertise remains crucial in the later phases where precision, usability, and technical quality are essential. This study contributes to the evolving conversation around human–AI collaboration, offering practical insights into how designers can work more effectively with emerging generative technologies.
The study investigates how tools such as ChatGPT, Adobe Firefly, and AI-based 3D model generators (like Meshy) can be applied across the design workflow. In the early phases, ChatGPT is used to simulate product benchmarking and create a user persona based on Finnish market expectations. Adobe Firefly is used to generate a visual moodboard aligned with that persona. A survey is also conducted to gather insights from design students and professionals on how AI is perceived in the field. In the practical component, an outdoor light post concept is developed using generative AI tools, while a parallel version is modeled manually in Blender. This enables a comparative analysis in terms of time efficiency, mesh quality, and editability.
By analyzing both research and modeling outcomes, the thesis highlights the strengths and limitations of current AI tools in industrial design. It argues that while AI can meaningfully accelerate and inspire early-stage tasks, manual expertise remains crucial in the later phases where precision, usability, and technical quality are essential. This study contributes to the evolving conversation around human–AI collaboration, offering practical insights into how designers can work more effectively with emerging generative technologies.