![\"River's](\"https:\/\/www.newsbeep.com\/au\/wp-content\/uploads\/2026\/06\/e513e922e658a5f5fc052c52a178b47f.jpg\")
<\/p>\nSian Willmott: There\u2019s a lot of energy in the industry right now around how digital tools are starting to reshape the way we design and deliver projects. It feels like we\u2019re moving past this experimentation phase into something a little bit more embedded in our everyday practices, particularly when it comes to things like documentation, embedding sustainability requirements for procurement and data collection to support decision-making.<\/p>\n
At Hassell, we have a few different areas of focus when it comes to digital integration. One of those is materials. We\u2019ve been working with a UK start-up called Materials Index to conduct artificial-intelligence-enabled audits of materials, and fixtures, fittings and equipment on some of our projects. That helps us understand things like weight, carbon and ultimately the reuse potential of existing materials and products. We\u2019ve also been working with a platform called Materials Council on a digital materials library to help us with more informed material selection.<\/p>\n
Are either of you seeing digital tools start to reshape how materials are specified, evaluated and reused, or is your focus with artificial intelligence (AI) currently elsewhere?<\/p>\n
James Loder: As we started to delve into AI in the past three years, we had a front-end focus \u2013 exploring how we could assist our early conceptual phases with various image-generation tools, like Midjourney and Stable Diffusion. It was a lot of fun and there\u2019s still a place for supporting conceptual exploration in the early phases of a project, however, we\u2019ve quickly realised that AI isn\u2019t here really to design a building for us. It has the potential to provide us with access to information and data analysis to help us make more informed decisions, and to have more time for creative discussions.<\/p>\n
It\u2019s been interesting as a mid-sized practice to think about where we invest time in enabling AI. There\u2019s been a rush for many practices thinking they\u2019ve got to go and build their own large language model (LLM) with ChatGPT or Claude. Perhaps small- and medium-sized practices don\u2019t have the resources to invest in that type of tool, especially when the technology is moving so fast. Anything you think is the right solution is probably an outdated solution in a few months\u2019 time. Even this article will describe a particular moment in a rapidly changing environment.<\/p>\n
We\u2019re taking a nimble approach that builds AI literacy across the practice, enabling our team to think about how to engage with it and to be creative with making their own tools that provide data and information analysis to support our understanding about various materials and systems, and assist us in our life cycle assessment (LCA) process, earlier, as we\u2019re making design decisions.<\/p>\n
<\/p>\n
View gallery<\/p>\n
SW: Josh, what about at Archier?<\/p>\n
Josh FitzGerald: As with most companies, a few years ago we experimented with AI to draft emails and as a replacement for basic web searches. We were finding AI would lead us down a somewhat generic route and we had to spend considerable time to fact check and reinstate our personal touch, which, being a small practice, is often the reason we\u2019re engaged.<\/p>\n
As a carbon-neutral office, we\u2019re also constantly looking into the environmental impact of the tools and processes we use. A lot of people don\u2019t appreciate the impact that AI has relative to something like a basic Google search: it uses 10 to 15 times the amount of power.<\/p>\n
As a result, we started engaging with AI to build tools within the office that add value for the client, ensuring there is a net-positive outcome. One of the first we tried was fee proposals. Automating fee proposals only saves a few hours per project, which you would never be able to justify financially, if you were to engage a specialist to develop for you. However, using AI-assisted coding tools, we\u2019ve turned this unexciting \u2013 yet, important \u2013 document into an engaging, dynamic, interactive process for our clients.<\/p>\n
Now we have a personalised, dynamically generated project scale and scope, we can use the same tool to predict the timeframe for us to design and develop that project. It also generates a Gantt chart \u2013 a stage by stage breakdown \u2013 customised for our client and their project.<\/p>\n
![\"Yandoit](\"https:\/\/www.newsbeep.com\/au\/wp-content\/uploads\/2026\/06\/57772951fffe501e4e2f0e2b67cb4222.jpg\")
<\/p>\n
View gallery<\/p>\n
Back to the idea of environmental impact, this is one example of how, instead of using AI on an ongoing basis, we\u2019ve used it once-off to build a tool that allows us to have more meaningful conversations with our clients and which we can use to uncover misunderstandings or misalignments. Other tools that we have created \u2013 that generate planning reports or that allow us to quickly prepare a cost estimate at each stage of the design process, for example \u2013 allow us to be more present, so we can focus on getting to know the design better. We\u2019re careful not to use it to output anything client-facing, but we\u2019re trying to make sure there\u2019s always a net positive gain.<\/p>\n
JL: Josh\u2019s point there is a really good one \u2013 AI has a footprint. It is not well known enough by anybody using it to make a critical decision. As people adopt these tools, they need to be far more conscious about it. AI can actually help to create a tool that tells us what the energy use of any given prompt or command will take, so we can be more aware as we experiment.<\/p>\n
JF: A semi-scary fact is that over the last year across Archier and Candour, our energy consumption from using AI to build tools is equivalent to about two months\u2019 power use from the house I live in.<\/p>\n
SW: It\u2019s good to have that value judgment over what\u2019s strategic and what\u2019s not.<\/p>\n
![\"Arup](\"https:\/\/www.newsbeep.com\/au\/wp-content\/uploads\/2026\/06\/799f2eccda5ca529cbe4c724038e36b7.jpg\")
<\/p>\n
View gallery<\/p>\n
Interestingly, based on what you said earlier, James, Hassell is a really big practice and we have our own LLM \u2013 effectively a controlled version of ChatGPT, integrated with things like Google Gemini\u2019s Nano Banana and Veo for image and video generation, respectively \u2013 and we\u2019re also a carbon-neutral practice and B Corp certified, so there\u2019s an interesting tension there.<\/p>\n
As an internal system, our LLM allows us to plug and play all of our policies, reports and fee proposals. It also allows us a secure environment to build some of those custom tools that you\u2019re both talking about, to help our designers with the delivery of their projects where there\u2019s value doing so \u2013 like cutting out day-to-day, menial tasks that so we can focus on the creative or strategic delivery of our projects.<\/p>\n
JL: It\u2019s a creative act in itself to think about the tools that we need and how to use them.<\/p>\n
There was this flurry of third-party tools that flooded the market. Monthly subscriptions made them accessible for small practices, but because their user data is going to a third-party operator that then taps into one of the larger models, like Nano Banana or Claude, they\u2019re less secure.<\/p>\n
Now, we can make our own tools that are directly based on Claude or ChatGPT and it\u2019s far more secure, it\u2019s far more reliable, it\u2019s tailored and you\u2019re not paying for it. While we no longer need a technical person in the office, we do need someone or a group defining the culture, the ethics and the guardrails around AI use. And that\u2019s constantly shifting \u2013 no policy will stay the same.<\/p>\n
If you imagine the AI tool is like an expert sitting next to you from which you can seek guidance or advice, it\u2019s an important reminder that, at the end of the day, no one\u2019s going be able to hold that AI tool responsible. Unlike a specialist consultant who has their own liabilities and insurances, AI has no accountability, and it comes down to our interpretation of what we\u2019re given.<\/p>\n
![\"Berrbang](\"https:\/\/www.newsbeep.com\/au\/wp-content\/uploads\/2026\/06\/25d8c33e44565819020912740c35aae4.jpg\")
<\/p>\n
View gallery<\/p>\n
JF: That\u2019s a really good way of putting it \u2013 as soon as you generate something and send it, you are now responsible.<\/p>\n
There\u2019s the argument around AI stealing jobs, but on the flipside, we could never afford to hire someone who could build the tools AI is helping us with \u2013 it is augmenting our imagination and needs.<\/p>\n
SW: The social dilemma is interesting, isn\u2019t it? I wonder if we were having these same conversations when computers became technology for all of us to use, or when we all got our hands on an Excel spreadsheet and could suddenly do complex maths, or when Revit came out and we were able to build coordinated, data-rich 3D models? They\u2019ve all been leaps in technology that have enabled us to do things quicker, and allowed us to give more time to the core problem solving aspect of our jobs.<\/p>\n
Have either of you seen learning, consistency and confidence within your teams shift with AI?<\/p>\n
JF: My first thought here is that confidence and competency are two slightly different things. I\u2019m seeing people be more confident because they\u2019re exploring new things and they\u2019re finding AI can give them a leg up when they\u2019re in unfamiliar territory.<\/p>\n
Conversely, I have a concern competency is dropping. On your idea of technological leaps, Sian, I think AI is analogous to the advent of computer-aided drawing (CAD). CAD greatly expanded what we could produce as architects, however there is an argument it came at the cost of understanding the intention behind what was being produced. AI will likely follow this pattern: an expansion of capability and production, at the cost of personalisation and intention.<\/p>\n
In that sense, there\u2019s been a skill increase in one area but decrease elsewhere.<\/p>\n
![\"Off](\"https:\/\/www.newsbeep.com\/au\/wp-content\/uploads\/2026\/06\/846dbf67d1e4f54682fc0b61c56b90b6.jpg\")
<\/p>\n
View gallery<\/p>\n
JL: With each technological leap, worlds continue to become more complex, more process driven, more regulated. In a funny way, we gain and then lose the time, and we actually haven\u2019t really gained more time.<\/p>\n
We\u2019re at another leap. AI is the thing that will help us to assess, process and analyse, but it won\u2019t replace the more tactile drawing act that CAD did. The front end is still the creative human act.<\/p>\n
Confidence is low because AI was fairly unreliable at first. That\u2019s why people are still hesitant, because fundamentally they don\u2019t want to be held responsible. We did a survey recently and we had it was 85 percent of our of our team using AI, which seems like a good uptake, but there\u2019s 15 percent who are currently not engaged with it at all. There\u2019s some people who are really pushing it creatively, and in the next few months we\u2019re looking to run workshops to try and reframe people\u2019s mindset about how to be more exploratory, and build both the confidence and competency in AI use. <\/p>\n
SW: Looking into the future, what\u2019s the one thing the industry is still doing manually today that you think will feel completely outdated in just a few years or months?<\/p>\n
JF: AI is not this thing that\u2019s going to take over the world, nor fix everything for the better. It comes at a cost. We need to be mindful of that and use it respectfully.<\/p>\n
I\u2019m not an evangelist of AI, but the people that are refusing to use it \u2013 who likely have their reasons \u2013 I feel they\u2019ll be left behind. AI is moving fast, in a really exciting way. What you can do with AI-assisted coding tools is literally coming down to your own imagination and creativity. It\u2019s no longer a technical concern. If you can dream of something, you can build it. And that is unbelievably exciting; it\u2019s also daunting as hell.<\/p>\n
![\"First](\"https:\/\/www.newsbeep.com\/au\/wp-content\/uploads\/2026\/06\/dfa8cc7ea4151af2903090f7dafe1114.jpg\")
<\/p>\n
View gallery<\/p>\n
JL: I\u2019d say the current biggest gap for the building industry is access to current, relevant data. In the tech industry, information on software is online and accessible by LLMs. But in architecture, our outputs are physical. We move from designing one building and onto designing the next, without an adequate assessment of the buildings we\u2019ve built.<\/p>\n
Where\u2019s the post-occupancy data on the performance and operation of the building, and also the way its spaces are used? How can we, or our clients, capture and harness that data \u2013 and it\u2019s really compelling data \u2013 to shape briefs for a new building that might be smaller than what we think we need? Likewise, the ability to have more readily accessible construction cost information as we design would be very powerful, and would save a lot of wasted time and money in the industry.<\/p>\n
JF: I love the idea that AI can help us build tools that allow us to engage with, and be more accountable for, the whole design process. As you just mentioned James, construction costs are key to the design process, however we\u2019ve never had the ability to meaningfully engage with them \u2013 resulting in wasted time and effort. We can now build tools to capture and action this data.<\/p>\n
\u2013 James Loder is a partner and design director at Wardle, leading design across the practice, from single homes to public, education, cultural and civic buildings. He cultivates Wardle\u2019s design culture, developing architecture attuned to the people who use it, Country and place, and environmental responsibility. He also mentors emerging designers and leads Wardle\u2019s use of digital tools and technology-led design, extending the practice\u2019s longstanding curiosity about new forms of making.<\/p>\n
\u2013 Josh FitzGerald is a founding director of Archier. With a deep understanding of how buildings come together, FitzGerald brings a process-driven approach to design and delivery, balancing technical precision with a strong focus on collaboration. He has been instrumental in shaping Archier\u2019s BIM strategy, enabling clear communication, efficient coordination, and seamless integration across project teams. FitzGerald is also the founder of Candour, a panelised construction system that simplifies the process of delivering sustainable, high-performance buildings.<\/p>\n
\u2013 Sian Willmott is a sustainable design leader at Hassell, with expertise in systems thinking, regenerative design and circular materials. She works across architecture, interiors, landscape and urban design, using integrated and interdisciplinary approaches to uncover opportunities for additional environmental, social and commercial value aligned with stakeholder drivers. Her work focuses on connecting strategy, design and delivery \u2013 helping clients navigate complex challenges through practical, scalable solutions that strengthen long-term resilience, performance and project outcomes.<\/p>\n