Neuron Case Study: 204 + 205 Fitzwilliam Street, Parramatta
September 22, 2023
Unlocking affordable housing incentives in NSW after recent policy amendments.
December 20, 2023
Neuron Case Study: 204 + 205 Fitzwilliam Street, Parramatta
September 22, 2023
Unlocking affordable housing incentives in NSW after recent policy amendments.
December 20, 2023

AI Keynote Speech:
Neuron Director Steven Cassells at the Property Council of Australia


Hello, thank you, it's wonderful to be here.

Let me start by saying despite running an engineering tech business, just like everyone in this room today, I too am on this digital journey trying to navigate this ever-changing digital landscape. After 20 years in construction, I can appreciate how tech and AI can seem a world away from our day-to-day life in property, so I want to start by bringing this back to the fundamentals; where we’ve come from, where we are heading and why I think it is crucial that the people in this room, both government and industry, get involved now.

Where we’ve come from.

The 1970s gave us the personal computer, a box that we could interface with, through zeros and ones. That gave us the building blocks of a new language that we could communicate with digitally. That, along with the internet, propelled us into half a century of building software to help us communicate, problem-solve, learn, get organised and do lots of wonderful (and not-so-wonderful) things, that have helped us progress as a civilisation. And it's incredible how far we have come.

It does, however, have its limits.

Today we pay software developers a lot of money to create apps in a language we can't speak, that do, (and this is important) exactly what we tell them to do. To improve it in any way requires more highly paid software developers to expand functionality.

This does two things;

1. It limits how quickly we can release new tools to fix problems we see in the world

2. It limits how complex a problem we can solve, as generally speaking the more complex, the more coding required, the more cost to build and maintain.

Typically the large tech companies are the only ones who can afford to have such tools, with huge workforces. And they generally focus on high-growth areas such as advertising or products that they can sell to the masses; email, task management tools and the like. Up until recently, few have attempted to penetrate our complex, obscure world of property!

Why AI has become so popular is because it finally gives the opportunity to build software to learn through exposure to data, we can now code software to self-learn (at least at a very basic level), attempting to mimic the human learning process and improving its performance over time.

This emerging new approach to building software will enable us to build less software that does more and tackle far more complex problems faster. This also creates more affordable ways to rapidly expand what problems we can now solve, including the complex problems we are faced with in property.

It’s like moving from a character on a computer game where we dictate every move and moving towards creating a child, a super smart child, and teaching it the rules of the world and letting it find its own way in life, within a controlled environment. It's really exciting! And the market is being flooded with innovative AI tools.

So where are we going?

The way I see it, we have a choice. As an industry, we can either ignore it, and try to live our lives not embracing the next wave of advancement, and wait to see what it turns out to be, or, we embrace it, just like we have done again and again with the computer, the internet, Microsoft, and Revit, and we drive where this goes with good people at the helm.

Fortunately, none of us need to know how to code or develop AI, there are tonnes of people who do, and they will sell us products that we can use far cheaper than creating them ourselves.

But the crucial point is this, the people in this room know the real problems we need to solve far more than the tech industry or coders writing AI apps. We really do need our industry experts to drive this process. If we provide a clear direction, AI and the tech industry will help us build the solutions.

So, if you agree, then, our next step is simply to choose which ones make sense for our businesses now and to start that journey of integration. For now, the tools are amazing but in comparison to what they will be, they’re quite straightforward. So, I want to give you some examples of real AI tools you could utilise now that can make an immediate impact on your business. I’ve chosen just three areas to explore to give you a taste. Feel free to jot these down and Google them afterwards.

This first area I would categorise as Risk Mitigation AI tools. Tools to help us make fewer mistakes.

These Risk Mitigation AI Tools include:

1. ChatGPT: The most famous one. It's incredible at finding information fast and helping assist with a wide range of tasks. Given it's only 8 months old, it's understandable that it's in its infancy and will not nail everything, so it's crucial you use it with care.

2. ZeroGPT: This AI tool enables you to drop in text and it will search for how much of that text was likely produced by ChatGPT. Schools use it extensively to check on essays etc It's a great way to sanity-check material.

3. Chat PDF: This is probably the most relevant tool to those in this room: it allows you to upload a PDF and ask it questions and it will search to compile an answer, kind of like a “find” command on steroids. Given the thousands of pages of reports we review, this tool could help really help.

Design Generation

This next category I call Design Generation. Tools that can rapidly generate ideas for a site.

Archistar:Who is here today, have a push-button solution that can generate concept ideas rapidly while comparing these against council planning regulations and a vast array of data sets. This process can be supplemented with your more traditional estate master and internal Excel sheets if desired, but it enables you to robustly test sites rapidly.

If you find one that looks good and you get an architect and planner involved to do a concept design, irrespective of what tool you or your architect have used, be it sketch-up, ArchiCAD, Revit or a cloud-based design tool, there are emerging AI rendering tools like Stable Diffusion and Mid-Journey that can take these preliminary ideas or envelopes and render them to look real, to help communicate your idea to a third party or council. These tools are a little tricky to use so probably best to put that task on an intern from your architect's office but these are incredible at converting loose ideas into realistic renders, fast.

These tools create a unique opportunity to collaborate with government and councils for early feedback on proposed developments. Currently, the absence of this step leads to months of design work and often wasted time and money that end in unforeseen consent issues, outright refusals, and feasibility problems. To address this, we need more input and analysis upfront, but developers can't afford this extra work, as they can't unlock more capital without certainty.

If councils would be open to utilising these renders and early concept tools to provide early in-principal feedback, that a design could be made approvable, developers could get more certainly, which will unlock more capital, to spend on refining concept designs knowing these improved design outcomes get them a step closer to DA approval. Waiting until post-DA approval for these key insights leads to compromised design outcomes across the board. Implementing this approach I believe will unlock my final category of AI and tech which I believe is the best solution we currently have to address Australia's property challenges.

Design Optimisation

We must figure out how to do more with less to address our current predicament.

Orbitalstack by RWDI, Lateral Consulting (ex-Meriton design leads) and my company Neuron- between us we now have tools that can provide rapid feedback on everything from utilities, wind, acoustics, structure (both substructure and superstructure), MEP services requirements that’s your substations, lifts, pumps, incl. cost data and visual feedback that can inform and refine a concept design. This isn’t new. Between us, we have supported hundreds of projects nationally helping each of them find what optimum looks like for their site, including helping governments assess their own sites.

Optimising a design requires a holistic review of all data rapidly. This is what AI and tech do so well. This will help us refine towards shared desired project outcomes that are both commercially viable and align with the council's desires.


To finish, I mentioned at the start that I believe it is crucial that the people in this room, both government and industry, get involved now. So why now, and not 12 months from now?

Well right now, we have a major housing supply issue, we need well over a million homes to be delivered within the next 6 years to keep up with demand. If we look at our current planning system, right now a new project on average takes 4-6 years from inception to completion with only 1.5-2 years of that period being the actual construction. Some large precincts take far longer than this. Clearly, we have an almost 0% chance of meeting this supply with our current processes and timelines.

We know the capital is out there if we can provide certainty of returns and time frames. If we don’t, the capital goes elsewhere, and we don't get new buildings.

So we have a certainty and speed problem. That's our Achilles heel.

From a government perspective, our planning system was never designed to scale. This is a core issue, deeper than short-term policy changes or incentives.

From a construction perspective, it also cannot scale. There is a finite number of people available to build. This means we cannot assume we can back-end the delivery of new building stock in years to come.

We must find a way to steadily increase supply immediately, and then continue to ramp back up sustainably.

As an engineer who works on literally hundreds of DA submissions, looking in from the sidelines, I feel better data and insights can help inform better conversations. To educate both parties to help understand the others' position, improving trust between developers and councils. This is critical, as I can't see how we get out of this situation without compromises. We can’t have it all.

And without trust from both parties, conversations on compromise cannot occur.

Tech and AI can fill this gap. It can be a mediator or a bedrock to fast-track decision-making. Automate the mundane at scale and free up resources to sort out complicated issues with trusted independent data available.

The consequences of us failing are severe, once rent rises above 30% of a person's salary, their lifestyle starts to deteriorate rapidly; Rental properties will become filled with mattresses as people cram in to cover rent, similar to Vancouver when I lived there. Young Australians will grow up to learn what their personal wealth looks like without having equity in a home and what it's like to retire as renters.

In Ireland, post-GFC, homelessness, suicide, depression and crime also rose as people became increasingly desperate to keep up with the cost of living.

It doesn't have to be this way if we act now.

I feel highly motivated and excited at the prospect of trying to solve for scale within the planning system. I feel a responsibility to continue to upskill, so I can do what I can to improve solutions and processes within my sphere of influence.

We shouldn't confine our industry to the limits of our present imagination, it's been done so many times before.

So, be curious, be cautious, be brave. Because the right people, using the right tools, will unlock a world of change.

Thank you.