However, AI’s capabilities extend well beyond visualisation, boosting efficiency and productivity to revolutionise many aspects of the business and practice of architecture. Regardless of the size of a practice, AI can be harnessed to streamline design and delivery, freeing up time for the creative work that only humans are capable of.

Speed up information gathering 

Off the shelf, web-based AI products such as ChatGPT and Copilot can search the web for information to assist with tasks such as generating text for reports and submissions or be trained on specific sources.

‘Training’ an AI tool means telling it where to look for something. An example is a ChatGPT assistant I have trained on the National Construction Code (NCC) standards. Ask: “I’m designing a stair for a Class 9 building with a floor-to-ceiling height of 3200mm. What do I need to take into consideration?” The NCC assistant will instantly return the relevant information alongside links to the source, saving time and helping to ensure compliance.

This NCC ChatGPT assistant is freely available to anyone via this link.  

Mine your practice’s data and knowledge bank 

Off-the-shelf AI tools can be downloaded locally and trained on company-specific data to securely deliver tailored outputs. For example, train Microsoft Copilot on your internal business documents. Whether HR policies, project management protocols or quality assurance processes, Copilot can make these resources immediately accessible to all staff in real-time. This reduces delays and fosters a more informed
workforce.

More valuable, is the use of AI tools to leverage a practice’s knowledge bank. Working on a new sports facility? Ask the question, “What are the five biggest sports projects we have delivered?” The trained AI tool can find the projects, then go on to provide specific information on each. This is a revolutionary way to make the years of experience and expertise embedded in a practice accessible to all staff. AI visualisation tools can also generate images using only the practice’s projects rather than drawing on unfiltered content.

The best way to deploy safe, secure, and cost-effective AI to access digital assets is to download Large Language Models (LLMs), train them on proprietary business data and run them locally within your secure IT environment. This offers a higher level of security than a web-based product, enabling staff to interact with the AI tool through a custom-built locally installed desktop user interface.

Streamline documentation and reduce risk 

AI tools can greatly assist with the efficient creation of accurate documentation. They can automate the generation of technical drawings and specifications and be linked to REVIT families, speeding up repetitive tasks. They can tailor design recommendations to specific client preferences and requirements.

AI can perform building code analysis and compliance checks and automate defects reporting and ordering. This has great potential for reducing the risk that architects are increasingly burdened with by feeding data directly into BIM models to share responsibility more equitably with consultants and contractors.

What to look out for – lessons learnt 

AI is a tool and like any technology, it has challenges and limitations. All work and data should be checked by a human – people remain responsible for what is created, not bots. Ensure your AI assistant is designed to be transparent in its decision-making processes so that architects understand how it arrives at its answers and trust its recommendations. For example, including citations.

Your outputs are reliant on the quality of the data your AI tool is trained on. Ensure your data is high-quality, relevant, and up to date. Be vigilant about potential biases in your AI models. The tool may be trained on projects that have a different brief to the one you are working on, for example, the climate, the dominant transport mode, the materials selection, or aesthetics.

Prioritise robust data protection measures to safeguard sensitive client information and intellectual property. Be aware that AI is an energy and resource intensive tool. High-level data processing and image
generation uses large amounts of power and water, which has ESG implications. I anticipate that there will soon be tools that can quantify AI’s environmental impact. All technology needs to be used responsibly – don’t use AI when there’s something else that can do the job, like a simple calculator.

Embrace AI

AI technology has fast become integral to the practice of architecture, as complementary to human creativity and expertise. Our industry has a great opportunity to reshape the way we work by the thoughtful application of emerging technologies. It is critical for architects to take an active role in developing the AI tools that will help to shape the built environment of the future – one where intelligent data use through robust AI implementation will define success. Embrace AI and reap the benefits.

Architecture Bulletin February 2025

📷 Jes Lindsay