AI has moved quickly from provocation to practice, and the conversation among architects is shifting. The debate is no longer centered on whether artificial intelligence will influence architectural design and visualization — the answer to that is a definitive “yes”. The more pressing question is how architects can integrate AI into practice without weakening the pillars of judgment, authorship and technical understanding that define good design.
That question sat at the center of our recent webinar with Roderick Bates, Senior Director of Product Operations at Chaos, who joined Editor in Chief Paul Keskeys to unpack the findings from the new Architizer x Chaos report, How AI Is Reshaping Architectural Design & Visualization in 2026. The report, based on responses from nearly 800 architects and designers worldwide, paints a picture of a profession moving steadily toward broader AI adoption. AI remains experimental for many firms, but momentum is growing.
The numbers reveal a profession finding its footing: around 60% of firms surveyed are now actively using AI in some capacity, while 86% of current AI users report saving time through these tools. Enthusiasm for increasing AI use over the next year is remarkably high. Yet throughout the webinar, Bates repeatedly returned to a larger idea: architects are beginning to distinguish between what AI does well and where human expertise still matters most.
AI: An Enthusiastic Assistant
To explain that relationship, he reached for a comparison that reframes the AI narrative. Bates described AI as “the overeager intern,” noting that “sometimes that overeager intern does fantastic work, really impresses you, and sometimes you need to tell them to slow down a little bit and think a little bit harder.” This grounded metaphor feels apt as it avoids the extremes that often dominate discussions around AI, leaving room for enthusiasm without surrendering to hype.
Architects, in Bates’ view, are increasingly discovering that successful use of AI depends less on prompts and more on judgment. Survey findings support that position. Satisfaction with AI remains mixed, and concerns around quality and reliability continue to appear year after year. AI can save time and accelerate workflows, but few architects appear ready to let it run without oversight.
Chaos AI-powered Veras software can now be used inside mainstream modeling applications such as Sketchup, turning 3D views into fully rendered, photorealistic images in seconds. Image courtesy of Chaos.
Bates was direct on this point. If AI is the overeager intern, someone still has to supervise it. “You need to be the critic to that intern,” he said during a discussion on architectural education and emerging professionals. The comment surfaced in response to a concern many architects are wrestling with: if AI begins automating tasks that traditionally helped younger designers build expertise, where does experience come from?
His answer pointed back toward the foundations of architectural education. Schools, he argues, may need to place renewed emphasis on design fundamentals rather than continually chasing new technologies. Understanding composition, materials, spatial relationships and building performance becomes increasingly important when designers are expected to evaluate machine-generated outputs critically. As Bates put it, “the fundamentals are always the fundamentals. It’s not like AI is changing what the building is.”
That observation cuts against some of the louder narratives surrounding AI. Technology may alter workflow structures and compress production timelines, but architecture itself remains architecture. Questions around light, comfort and human experience still sit at the center of practice.
Reversing the Rendering Workflow
Elsewhere in the webinar, Bates highlighted examples where AI is beginning to reshape workflows in more tangible ways. He demonstrated experimental hybrid workflows where AI-generated lighting layouts can move into editable geometry and then back into tools like Enscape for refinement and visualization. Rather than replacing established software ecosystems, these workflows suggest a future where AI reduces friction between design ideas and visual communication.
Chaos’ AI-powered Morphus tool enables architects to use prompts to add details to their models, such as lighting layouts, directly within Revit. Image courtesy of Chaos.
The most provocative example involved a reversal of a familiar process. Bates described customers who once modeled geometry first and rendered afterward, but are now moving in the opposite direction. Some designers, he explained, are beginning with AI-generated imagery and using those outputs as the basis for subsequent design development. In his words, “AI rendering almost takes on the role that sketching used to have.”
This shift feels significant. Sketching has always held value because of its speed and looseness. It allows ideas to emerge without demanding precision too early. AI-generated imagery appears capable of serving a similar function, helping architects explore spatial directions before committing significant time to modeling and production.
From Fear to “Figuring It Out”
Perhaps the webinar’s strongest takeaway was that the conversation around AI is maturing. Earlier discussions often focused on fear — fear of replacement, of diminished creativity, or of losing control. Those concerns have not disappeared, but they increasingly sit alongside more practical questions: Where does AI save meaningful time? Which workflows benefit most? How can architects integrate it responsibly?
According to the survey findings, many firms appear willing to answer those questions through experimentation, putting ideological concerns to one side as they do so.
Key findings from the survey reveal increasing adoption and an evolving mindset on AI in architectural practice. Image courtesy of Chaos.
According to Bates, the profession is not handing architecture over to machines — it is learning how to direct them, while retaining the control that ensures high-quality projects are the end result. In fact, as AI becomes more capable, this may mean the architect’s role becomes even more important. After all, as Bates reminded attendees near the end of the webinar, “you still need to own the outcome.”
To learn more and download your free copy of the report, click here, and learn more about Chaos’s latest developments in architectural visualization here.
