Mastering the machine
The potential of AI for architecture is too riveting to shy away from.
Picturing the future
How AI halved the time it took us to visualise the future of Brisbane and SEQ
For designers participating in our annual Charrette event, the speculative design challenge is a perfect opportunity to test new tools and approaches that can help us picture the future.
This year our teams exclusively used AI to visualise our final concepts for a more climate-responsive South East Queensland – a process that halved the time it typically takes to produce realistic renders.
This in-house, AI-powered approach gave us greater control over our creative output while teaching us valuable lessons about how to balance digital innovation with human ingenuity and insight.
Growth and opportunity
Like so many industries, AI is gaining ground in design, sparking heated debate about whether its use enhances – or undermines – what we do.
In Australia, 73% of professionals working in the built environment view AI as an opportunity – well above the global average of 60% – and around 85% are using it at least once a week, according to a 2025 Arup survey.
Visualisation appears to be one of the most widely used applications of AI in our field.
But that comes with challenges, something that was clear when translating our future-focused ideas into faithful imagery.
“AI’s real value lies in how we use it,” says Architectus CEO Ray Brown. “The quality of what you get back depends entirely on the quality of what you put into your prompts.”
“Designers will always need a deep understanding of the site and context as well as technical knowledge and experience. AI just helps us express those ideas faster and more powerfully.”
— Architectus CEO Ray Brown
The power of the prompt
For our 2024 Charrette we setup a dedicated team of digital experts focused on AI and automation. This time around, we put more tools into the hands of participating designers, who explored ways to enhance rather than replace the human element of design.
While some teams successfully incorporated AI in the ideation and design critique phases, the real focus was post-production, using programs such as AI video generator Google Flow and text-to-image tool Nano Banana, part of Google Gemini.
For example, architects could prompt with their hand sketches, turning them into evocative and realistic 3D images or video while still maintaining creative control and design integrity. Other scenarios included feeding in screenshots of 3D Rhino models as well as specific, descriptive text prompts.
Designers involved in the visualisation phase can see both sides of the argument around AI use.
On the one hand, it’s difficult to avoid AI’s tendency to ‘smooth everything out’, making buildings and places more generic, they point out.
There’s also an element of ‘rolling the dice’ with AI. No matter how carefully we prompt, it often returns unexpected results that aren’t very useful.
But AI’s speed and ability to deal with complexity can make designers more nimble – a major benefit this year considering the scale of the SEQ corridor and the number of sites to model.
One of the biggest upsides may be AI’s potential to expand our horizons, allowing us to experiment with typologies and worlds limited only by our imaginations.
