By Tracy Hanes
Artificial intelligence is making a dramatic impact on the home-building industry.
How is artificial intelligence (AI) being utilized in the building industry?”
It seemed a logical starting point as I typed my question into ChatGPT. Created by San Francisco-based OpenAI, ChatGPT is a free-to-use AI chatbot that has existed since late 2022, has 180.5 million users worldwide, and automates text-based tasks. It uses large amounts of data and computing power to string words together in an understandable way.
The answer ChatGPT provided was long and detailed, but here are the highlights: In the building industry, AI is being used for such things as planning and designing projects, construction management, safety monitoring, material management, energy efficiency and sustainability targets, and quality control—and it can complete most tasks with lightning speed.
A lot of the foundational work for AI has been done in Canada. Toronto has the highest concentration of AI start-ups in the world. Montreal is another hub, where the company Mila – Quebec AI Institute has the largest concentration of university researchers in the “deep learning” field in one place. Mila is a collaboration between Université de Montréal and McGill University, in close cooperation with Polytechnique Montréal and HEC Montréal. Mila and sister institute Amii (based at the University of Alberta) and Vector (based at the University of Toronto) play a central role in the first national AI strategy in the world—the Pan-Canadian Artificial Intelligence Strategy—led by the Canadian Institute for Advanced Research, a global research organization.
As dazzling and disruptive as it is, AI still needs human input—for now, anyway.
So, what’s the human take on AI’s current and future role in the home building industry? “We believe this technology is a game-changer and phenomenal in what it does. I can see the future of architecture while designing with it,” says Kostika Lala, founder and principal architect of Toronto-based LALA Studio, a full-service architecture and design firm.
Lala’s studio has been exploring the possibilities that arise from integrating AI with architecture. “It’s so good and it saves so much time. When CAD (Computer-Assisted Design) came in, it threw the industry upside down, and everyone had to go out and get computers and learn the software. I see a similar phenomenon as a result of this.”
Lala is also co-founder of AI rendering engine Davinte, a Toronto-based start-up looking to integrate its AI capabilities into the mainstream architecture industry. Davinte is working with some of the world’s largest CAD and BIM software firms and hopes to bring its software plugin to users early this year.
The new kid on the block is generative AI, notes Frank Magliocco, Canadian Real Estate Leader for PwC Canada. “We’re at the front end of the trend. It’s very broad and it’s moving very fast.” Traditional AI is nothing new (think Siri or Alexa). It responds to a particular series of input data. For example, if you play computer chess, the computer predicts its opponents’ moves based on strategies or rules it was programmed with. Generative AI learns from input data to create something new.
In a September webinar, AI and Housing Needs Assessments, Matt Parker, executive vice-president of data and analytics for Calgary-based HelpSeeker Technologies, said that while AI has been around since the 2000s, there’s been a recent explosion in its use due to technological improvements that make it more accessible than ever. Technology that used to be difficult to interact with and cost millions of dollars to run has evolved into tools such as ChatGPT that provide free information within seconds.
However, “in terms of disruption, it’s no secret that real estate companies are laggers when it comes to adopting new technology,” says Magliocco.
There are a few exceptions in Ontario’s home-building industry, however. One is the Vaughan-based Plus Group, which is comprised of the companies RN Design, SRN Architects, SaleFish Software, Studio Uno and Coolaid Studios. SaleFish was developed as a platform that leverages tools and software to sell homes more effectively. It came out of the need in the low-rise industry to deal with selling large sites and allowed salespeople to see in real-time what lots had been sold, what home styles could go on a particular lot, etc. It was adapted for the condo market and evolved into an online sales platform during COVID-19.
Mike Robinson, chief technology officer at the Plus Group, sees AI as a valuable tool for its companies. “ChatGPT can free up lower-end labour and free people up to do what they were trained to do. When the photocopier came into offices, it freed people from copying to do less mundane work. For example, Robinson suggests GPT can be used to draw up contracts and agreements and ensure they are correctly written, or to compose client emails.
Lala says that while design is a focal point for his studio, it can also be used to achieve objectives such as structural performance, energy efficiency and aesthetic preferences. As well as integrating with Building Information Modeling (BIM) technology, AI can also be mobile and portable to provide on-demand design help.
Lala says it will aid with quality control, creating collaborations and designing better—and more distinctive—buildings. There are a lot of similar-looking buildings in Toronto because they are efficient to construct and the costs are known, Lala explains. Clients are hesitant to take on something unique if the costs are uncertain. However, AI technology can specify a project’s cost and other considerations.
“Humans will still create unique designs without GPT, but it can help improve it,” he says. “What I’ve done with OpenAI is take it and build our own technology, our own GPTs. That’s where the fun begins.”
VISUALIZING THE FUTURE
Magliocco says generative AI has many applications, from creating visual property mock-ups of developments before they are built, thus reducing costs, to producing content for marketing materials. One current application is predictive maintenance analysis, which takes data from a building, generates wear patterns, and addresses maintenance issues, such as when elevators need to be repaired.
AI is already being used to plan electrical and plumbing system routing. From a worksite safety standpoint, supervisors are being alerted to current and potential safety issues and construction errors while tracking workers, machinery and objects.
It can also be used for ESG (Environmental, Social and Governance) to simulate potential changes to land use, such as what happens if sea levels rise—something that’s particularly useful to the insurance industry, says Magliocco. It can be applied to zoning simulations to see the impact on property value. Or you can use it to screen tenant applications to determine suitable candidates, analyze mortgage applications and decide which ones are at risk of default.
Lala says it could also assist in streamlining the approvals process, as AI could analyze basic building drawings before they go to municipal planning staff.
HelpSeekers uses AI to help produce housing assessments. The pluses have been speed and scalability, says company co-founder Dr. Alina Turner. The technology can handle vast datasets and produce reports without human error. For example, in a region of 30 very diverse municipalities, including cities, hamlets and villages, the AI tools can quickly provide multiple assessments within hours, fully customized to each municipality.
Turner gave the examples of Brantford and St. Catharines, two nearby Ontario communities experiencing growth. Brantford has a large swath of greenfield that can be developed into single-family homes, while St. Catharines is mainly brownfield development and needs density added. Even though they are in the same jurisdiction, the two communities are very different, and AI can help generate assessments that reflect the different approaches they’ll need as they grow, such as the required infrastructure.
Turner cited another housing assessment done using AI for Edwardsburgh/Cardinal, a township in eastern Ontario south of Ottawa. The community has less than 10,000 people in various hamlets and villages. HelpSeekers had to dig into the neighbourhood level to identify pockets of emerging core housing needs/poverty/social challenges. Turner says the township had never done a housing needs assessment before but can use the report to advocate with the province to get transitional supportive housing capacity in place before the issues become significant challenges.
“Even though this is AI-generated, combining the human expertise from us, as experts in the field, with their community expertise, is where the magic really happens,” says Turner.
Currently, Lala is working on garden villa designs using AI. Through large language models (LLM) technology, his firm can design homes and precisely visualize them on-site. The renderings for the villas demonstrate the power to generate designs using AI while controlling form, materials and weather conditions through AI prompt engineering. The application is still in the beta phase, but LALA Studio is collaborating with software engineers and machine learning architects to develop it further.
“Through my experience, I’ve done renderings and dealt with many natural habitats,” Lala says. “Let’s say you want to build a house on top of a mountain or on Mars. It’s very difficult to imagine what it can be like, but this can make it happen. It fills the gap between imagination and reality.”
Lala says AI can also track material availability. “As an industry, we can create a database where all manufacturers and suppliers list their stock and materials. You can allow AI to look through the database. If it says wood is cheap, we’ll go with wood for our garden villa. This is where the power of the technology really starts to take shape and helps us to build better buildings.”
SEEING IN BELIEVING
Robinson predicts AR and VR will be “very powerful in the future. It’s very good at conveying design ideas to customers and clients. We can look at plans and visualize them, as we’ve been doing this for 35 years. But customers can’t, and our clients have trouble with it. We have to do sketches or 3D renderings. What AI is doing is amazing. It can take sketches, plans and descriptions of the feeling we want to convey and render it.
“We use AI for quick internal visualization, but not for final work for clients yet,” Robinson adds. “It’s not quite to the quality we like for our clients. But it works well for internal discussions and quick ideas on different sketches or to look at possible finishes.”
Robinson says builders spend millions on model homes, but that won’t be necessary with virtual reality software. As well as seeing a home that’s been virtually staged, homebuyers will be able to walk through and see what different finishes look like. It is possible now, says Robinson, but his company hasn’t used it much yet for that.
Robinson has also played with digital twin technology that creates exact replicas of spaces, such as buildings under construction. The digital twin integrates real-time data from a building with its digital representation through BIM and 3D modelling software, sensors, surveying data, etc. It allows a construction team to get insight into every building component. For instance, they can see if a wall is out a few inches or if a plumbing stack is in the wrong place. It makes collaborations between team members easier and faster, creates more efficient workflows between contractors/trades, and reduces issues.
But as game-changing as AI can be, there is still much apprehension. “I think people in the market have a lot of fear,” says Robinson. “Is it going to take jobs or disrupt things? At the Plus Group, AI will be a tool like a computer, fax machine or cell phone. It’s going to take our work to the next level.”
Robinson says ChatGPT is impacting marketing in particular. It can do lower-end tasks well and quickly, but full campaigns still need people to fill in the gap, take the framework and add their own flavour to make a product stand out.
“Clients can do ChatGPT. They are paying us to do things they can’t do,” says Robinson. “They are paying for our experience and many years of knowledge. We can layer amazing work on top of the generic stuff ChatGPT does. For instance, we have a fantastic writer, and it reduces the time she has to spend on research. It allows her to shift her time to concentrate on the actual writing to make a better product.”
ONLY AS GOOD AS THE DATA
“It needs human input,” says Lala. “In a best-case scenario, it can handle mundane, routine tasks and scheduling and have them done immediately.” He says it can help with design tasks such as setting up lobbies, entrances and exits and building statistics. “These are very time-intensive tasks that are very important, yet not creative. I want people to create more beautiful things.”
“I think it’s going to create new jobs—tech jobs powered by human leads,” Magliocco says. “This technology can go through so much data so quickly, but you need the human touch as well. AI will eliminate jobs we don’t need to do, and people who lose those jobs will move up to where they can be more productive. Right now we have senior architects in the office doing manual tasks on a computer.”
Adoption of the technology is going to happen very fast, says Magliocco. “At PwC, we’re investing $200 million in generative AI and embedding it in all we do. It will have a profound effect on real estate.”
Lala says all big design/architectural firms are adding new departments dedicated to the technology. “The information is based on dataset design, and all buildings are done through data. The centre of it is data management and sorting and information sorting.”
However, adopting these new technologies must be done carefully and responsibly, Magliocco says. “You have to be systematic and not be chasing a lot of shiny objects.” The key is data quality and availability of data, he says. “If you have bad data, you will get bad information.”
Lala agrees. “The curation will be in the quality of data that the AI models have been trained with. Users will need to research, test and control what AI technology they are using.”
“AI tools are hugely powerful,” Parker notes in the HelpSeekers webinar. “It’s like a huge, powerful car that can go all these places and is convenient, but is still regulated and controlled. That’s similar to AI.”
It’s why adopting the technology requires companies to work with professionals to build proper governance, thinking about data quality and how it will be assessed, Magliocca says. That’s one of the services offered by PwC, which works with tech giants such as Amazon web services and Microsoft and their advisors.
“Some companies have data scientists in their shop. It can be very disruptive, but it’s like when we were introduced to the internet—it’s that level of change,” Magliocco says. “You don’t need to be leading the charge, but you have to think about how to use it to disrupt—to give you a competitive edge.”
Become a member of the Ontario Home Builders’ Association.