AI in Architecture – The Use Cases Impacting Projects and Operations

Discover how some of the world's leading architects use AI to enhance creativity, productivity, and efficiency.

Given its ubiquity in the news, you'd be forgiven for thinking that artificial intelligence (AI) is a new phenomenon. Yet, British logician and computer pioneer Alan Mathison Turing wrote the first successful AI program in 1951. So, what's changed – and what does it mean for you?

In a recent roundtable webinar hosted by Architects Journal (AJ) and Deltek, we unpacked the impact and opportunities of using AI in architecture. Drawing on feedback from AJ100 practices, AJ found that 44% of firms already use AI in some way, necessitated by the massive rise in data produced by connected digital technologies.

With the rapid pace of AI's development, now is the ideal time to assess the technology's potentially game-changing impact on your projects and operations. 

Read on to discover some of its most compelling use cases in architecture today, as shared by industry experts at Foster and Partners, Arka Works, Heatherwick Studio and Deltek.

Accelerate and Democratize Analytics

AI offers a powerful opportunity to turn your data into actionable insights. The kind that informs proactive decision-making and boosts process efficiency and productivity.

"We've been working on developing AI-powered apps that use predictive models to better visualize and contextualize our business data," says Martha Tsigkari, Senior Partner, Head of Applied Research and Development at Foster + Partners. "These models also offer real-time assistance that guides us through difficult problems, such as predicting how materials may react to different real-world scenarios like temperature and weather."

The practice's use of surrogate models trained on giant data sets to replace low-level analytical processes is more prevalent and dramatically speeds up its time to insight.

"Our surrogate models inform predictive analysis, enabling us to start playing with design geometry much faster," says Martha. "We previously had to wait hours or even days to perform the necessary analytics; now, it all happens in real-time. Plus, we can plug it into Rhino3D to further democratize access to AI-powered insights."

While its surrogate models are only 90-95% accurate, real-time feedback helps the practice make smarter decisions earlier in the design process, saving time otherwise spent correcting mistakes later. 

Disseminate Knowledge and Insight

The rise of connected digital technologies and solutions has created a sprawl of disparate data across many practices. Working in this way makes it difficult for teams to access data, like valuable project information and multimedia resources, and collaborate around it.

In response, Foster + Partners developed an AI-powered search tool that lets its architects use simple natural language queries to receive instant answers. It later adapted its AI tool to understand searches for drawings and PDF documents, enabling people to reference past project materials – anytime, anywhere.

"Knowledge dissemination is a huge part of AI's appeal to us, especially with the rise of large language models and multimodal foundation models," says Martha. "Now, we can easily reference archived materials without spending hours manually sifting through files or asking around to find that one person with niche project knowledge."

It's a use-case that Keir Regan-Alexander, Principal at Arka Works, is keen to accelerate across the industry to normalize fluid data queries and problem-solving.

"There's huge potential to train AI models and use them to serve up automated answers to routine queries," says Keir. "For example, we could simplify stage report writing by using AI to populate templates and calculate estimated fees automatically. And architects could use AI-powered chatbots to get instant answers to difficult queries that otherwise require them to navigate multiple platforms and portals."

Enhance and Automate Design Processes

A prevailing concern among many is that AI will replace human ingenuity and creativity. Still, as our experts are keen to clarify, it's actually a means to augment and enhance our natural strengths. 

"We've started developing applications that use machine learning for image generation as we want to augment rather than replace our designer's creativity," says Martha. "We've built apps that can make design suggestions directly in our 3D modelling platforms. And I expect we will soon be able to automate basic 3D model creation, helping our architects spend more time on acute, value-adding design tasks."

For Pablo Zamorano, Head of Geometry and Computational Design at Heatherwick Studio, generative AI is a chance to automate routine manual tasks and boost designer creativity.

"Our designers spend many hours each day performing simple tasks in Photoshop and rendering and analyzing models. These are prime candidates for an AI handover," says Pablo. "We experimented using generative AI to create detailed images, but they were too generic and hard to tailor. Instead, we use AI to automate the tasks that nobody wants to do – like generating and tiling assets."

At Arka Works, Keir is taking generative AI even further by running deterministic analysis to help improve project sustainability. This analysis lets his team tap into environmental insights and apply them directly to software from Autodesk.

"You can look at lightning, potential wind and other weather conditions using CAD plugins," says Keir. "It helps us identify new ways to reduce our carbon footprint and design and deliver sustainable projects that align with our ESG goals."

How to Manage a Responsible Shift to AI

However, you use AI, all our experts agree that one thing is vital: effective change management.

As with any emerging technology or cultural shift, people may initially resist or struggle to understand it if you don't show how it contributes to your overall business vision.

"AI will ultimately change the creative process, and the value in doing so must come from the top down," says Keir. "You need a strategic group to lead the plan so everyone knows how to use AI responsibly and effectively. Likewise, you can't afford to have people compromising data by using unsanctioned AI because you've been too slow to implement an AI policy."

Martha agrees and believes that a clear framework is needed to ensure that AI becomes a catalyst for innovation, not a source of fear or risk.

"Our ultimate goal is to create a workspace where we celebrate technology," she says. "AI can assist and enhance so much of the architectural process, but to make that happen, you need to educate and regulate internally."

Watch the Whole Discussion for More AI Insights

This blog covers a small sample of what you'll learn in the Technology in Practice: AI in Architecture webinar. Watch it on demand to hear more from our panel about how they are leveraging AI in their practices.