The design consultancy landscape is changing. Design timelines are shrinking, the demand for innovative yet efficient design is growing, and finding the talent to meet these challenges is more competitive than ever.
Our Architecture and Landscape team has embarked on an exciting journey by embracing a new suite of powerful tools enabled by artificial intelligence. Rather than replacing the creative spark that defines our work, these tools augment our creativity. This case study explores how we are leveraging AI to supercharge and optimise our design processes, foster innovation, and deliver better and faster results for our clients.
The Challenge: Time
Our traditional design workflow is linear (and iterative). It’s proven, but often time-intensive. Each step builds sequentially on the last; any significant change can send our designers back to the drawing board, consuming valuable time and resources.
AI-augmented design allows our designers to combine time-consuming steps into a single, fluid phase of creation. Our workflow becomes leaner and more responsive—we move nimbly between ideation and 3D rendering, with AI tools enabling us to visualise concepts as quickly as we can imagine them.
Case Study
The Brief: When tasked with designing a supporting facility for our client, we faced tight deadlines and the need for multiple design options for client review. Traditional methods would have required weeks of intensive work before presenting initial concepts.
Our AI-Powered Approach: Instead of spending weeks on detailed modelling for a handful of options, we turned to our AI toolkit. We fed the AI models a combination of inputs:
- Rough hand sketches capturing initial formal ideas.
- Simple 3D massing models to define scale and volume.
- A library of reference images for materials, lighting, and atmospheric mood.
- Descriptive text prompts like “a curvilinear, parametric facade with bronze fins inspired by the flow of water” to guide the aesthetic.
Using a suite of tools including Stable Diffusion, Comfy UI and Midjourney, we generated dozens of diverse, high-quality design variations. This process of “rapid design optioneering” allowed us to explore a vast creative landscape, testing different forms, materials, and lighting conditions almost instantly.
The Result: We delivered more options, faster, while maintaining the design quality our clients expect. The preferred concepts were then developed using our full suite of traditional and digital tools, but the AI-enabled front-end process saved countless hours and accelerated decision-making.
Our Expanding AI Toolkit
This project is just one example. Across the team, we are embedding AI into almost every stage of the design process, using a range of specialised programs to enhance our capabilities.
- Web UI and Comfy UI enable us to navigate and configure workflows across different checkpoints, Lora, and ControlNet
- Stable Diffusion 3.5 and Flux.1 enable custom image generation with precise control of the outputs
- Midjourney and Leonardo.AI excel at rapid conceptual visualisation and style exploration
- Cog Video X brings our designs to life through AI-generated animations
- CityEngine and Dynamo leverage procedural generation for urban design and parametric modelling
- Unreal Engine provides real-time rendering capabilities and first-person view spatial explorations
Creativity First: AI as a Co-pilot, Not the Pilot
It is crucial to understand that AI hasn’t replaced our design process—it’s amplified it. These tools are precisely that: tools. They require a clear design vision, careful prompt crafting, thoughtful sketches and meticulous direction to produce useful outputs. The creativity, spatial understanding, and design sensibility that define great architecture still come from human minds.
The quality of AI outputs directly correlates with the quality of human input. Poor prompts yield poor results; thoughtful, well-crafted inputs produce remarkable possibilities. Think of AI as a supremely talented assistant who can handle the laborious task of visualisation, freeing up our designers to focus on what they do best: solving problems and creating beautiful, meaningful spaces.
A Call to Innovate
The transformation we’ve experienced isn’t unique to our team—it’s available to everyone willing to explore. These technologies are evolving rapidly, with new capabilities emerging monthly. Whether you’re in architecture, engineering, planning or any creative field, there’s likely an AI application that could enhance your workflow.
I’d suggest starting with a single tool on a low-stakes project. Approach it with curiosity and familiarise yourself with its strengths and limitations. Our team has found these tools genuinely exciting to work with, and we think you might too!
About the Author
Alvin Shum
As the Technical Director at AIS Hong Kong, Alvin Shum leads the architecture and landscape architecture teams, drawing on over 18 years of extensive international experience. A registered architect and a member of the Royal Institute of British Architects, Alvin has a proven track record of successfully delivering large-scale, multidisciplinary projects, including the Hong Kong International Airport expansion, London Crossrail, Singapore MRT and the iconic Battersea Power Station redevelopment, as well as hospitality and mixed-use commercial developments such as Hari Hotel and Hann Resort.
Alvin earned his Master of Architecture in the UK and holds a Postgraduate Certificate in Built Environment Sustainability and Adaptation. Complementing his architectural expertise, he is currently an MBA candidate at Imperial College Business School, where he is focusing on corporate innovation, smart city developments, and the application of Big Data, AI, and Machine Learning.
He is particularly passionate about leveraging emerging technologies to reshape the future of the built environment. His current focus is on exploring and implementing generative AI and robotic process automation to optimise and advance architectural and landscape design, driving innovation and efficiency in every project he undertakes.
