A few people here have mentioned AI being useful for early iteration, but also raised the issue that the output is still just pixels, and that designers may lose control over shape or become dependent on the tool.
That gets at the question I’m most curious about:
How do people here see the balance between controlled 3D rendering and AI image generation evolving?
For some parts of the process, a generated image may be enough. Mood, styling, context, getting unstuck, quick form exploration. But in many industrial design workflows, especially once a product is moving toward CAD, CMF, stakeholder review, or launch assets, the visuals need to stay connected to the actual object: geometry, scale, materials, lighting, variants, and decisions that can be repeated or revised.
Do you see AI staying mostly in the early concept, sketch render, and moodboard stage?
Or do you think the more useful direction is AI inside the 3D workflow, helping with scene setup, materials, lighting, cameras, variants, presentation boards, and output management, while the designer keeps control over the actual product and visual decisions?
Full disclosure, I’m trying to build such a tool, combining the best of 3D and AI. The thing I’m especially interested in is agents that can help add functionality or operate parts of the workflow for you, right when you need it. I’m actively trying to understand whether this is moving in a useful direction or if designers see it differently.