How AI is changing site evaluation for architects

This past fall, in Kirkwood, Georgia, local developer Stryant announced plans for a “transformative redevelopment” on the former site of a popular barbecue restaurant. The half-acre parcel in this historic streetcar suburb east of Atlanta will soon be turned into 52 micro-apartments with a solar-shaded rooftop. 

The design plan for the site wasn’t the initial concept; a previous architect had pitched a 16-unit building with parking. But the developer decided to go with a proposal developed with Vitras.AI, a tool that’s helping developers and designers figure out new site plans and, in this case, maximize the profitability of the space. 

Patrick Chopson, AIA, a practicing architect as well as the co-founder and chief product officer of Cove, the firm that developed Vitras, created the tool to help architects and developers better understand zoning and building codes and show them what could be possible. Cove believes that by giving architects better insight into the economics of a project, they can more efficiently contribute to the final blueprint.

“It’s a game changer for the developer to know that the architect is a true partner,” said Chopson.

In the case of this 52-unit project, Vitras helped the architects find aspects of the code—for example, an exception to the transitional height plane rule, and a passage that allows mechanical penthouses in a stairway—that helped shape the final design. Factoring those regulations in with the local zero-parking ordinance allowed the inclusion of the additional units, made it feasible to develop more affordable units without requiring government subsidies, and greatly benefitted the developer’s bottom line and rate of return. Stryant submitted plans to the city in October and aims to break ground next summer, ahead of the initial schedule and with more units in the blueprint. 

“Cove was able to kind of cut through the knot,” said Stan Sugarman, co-founder of Stryant. “They came up with a design that allowed the lot to be developed by right, which helped it move a lot faster. That is definitely rocket fuel for a developer.”

Improving speed and efficiency 

Like many creative industries, the business of architecture is being impacted by artificial intelligence, which is concerning to a majority of architects, according to research released by AIA earlier this year. A number of new tools utilizing artificial intelligence, zoning and building codes, and land ownership databases have changed the speed and efficiency of the site selection process for real estate developers, and how developers decide what they want—and what’s possible—for their forthcoming projects.

A number of larger real estate brokerages such as Avison Young have their own in-house tools to help investors find where to buy and build. Software services such as Cherre, Prophetic, ReZone and Deepblocks have developed their own tools to make it simpler to figure out the best places to build. One tool, Pillar, constantly evaluates sites and sends developers favorable listings when new plots become available.

These types of tools also provide developers with the ability to identify how best to use these sites, helping them better utilize land purchases or maximize zoning rules to add additional units to a forthcoming development. They can help proscribe the dimensions of a project before design even begins. 

This additional intelligence impacts the way architects will interact with developers and clients going forward. Will these tools replace some of the consultation and evaluation work that typically includes architects, providing them with less work, or does it, as Cove’s Chopson argues, help them design projects that perform better for their owners? 

Chopson argues that by having the economics of the site at hand—being able to showcase how certain choices can help hit the yields developers want—tools like this give architects more freedom. 

“We have all the information at our disposal, and then we can step in as a great architecture team and design something cool,” said Chopson. “When you don't have the information at your disposal, you end up doing things that get thrown in the trash.” 

In the case of Stryant, this process encourages them to push to get more out of future sites, said Sugarman. He thinks tools like this will allow him and his partners to be more aggressive pursuing projects in neighborhoods where it might have seemed more complicated or costly to build. 

“This has definitely taught us that the right partner can allow us to exceed our expectations from the site,” he said. “I’m definitely going to push architects harder on the next project.”

The solar panels, in addition to electric vehicle charging, at the Kirkwood, Georgia development are an element that Chopson says Vitras helped make possible by making the code more clear and highlighting how the additional energy generation would impact utility costs and operating expenses. He’s used Vitras in multifamily developments to determine ways to design units that meet certain rent thresholds: for instance, by unlocking density bonuses that can help increase operating income for a project.

Maor Greenberg, co-founder and CEO of Spacial.AI, a tool that automatically creates engineering plans for buildings and evaluates code compliance, allowing designers to more rapidly figure out which designs work best, says his software isn’t trying to replace architects, just help them be more efficient. 

“It’s more about telling an architect what’s wrong instead of optimizing the layout,” he said.

Hitting the right numbers

Chopson pushes back on the idea that projects using Vitras exemplify value-engineering or water down the architect’s creativity or role. He said if AI alone designed a building, it would just end up with the most statistically median, boring potential outcome.

“As architects, we want to create a customized experience that meets the needs of the individual human that we're designing it for,” he said. “We recognize that nothing should be truly automated, and also nothing should be truly anti-AI, either.”

Designing with AI doesn’t have to remove the hand of the architect, argues Chopson. Drafting new building layouts, designing mixed-use projects for different retail tenants, and adding amenities or flourishes can all be incorporated into projects if they pencil out; if designers better understand, via AI, the financial metrics of success, they can meet them and present ways to incorporate their own ideas into a project that hits the right numbers. 

“We can't tell developers how to do their business, but what we can do is think of things they're not going to think of,” said Chopson.

Cove has been presenting design ideas to a mall developer in Ohio seeking to spruce up an underperforming commercial site. Chopson said their proposal, which included a ribbon-like ice skating path that threaded throughout the development, not only made the financial case for the proposal but also included stats about how the skate concept would improve performance by generating additional foot traffic. 

Do these AI-generated tools help architects work better with developers, or does it automate away some of the creativity of the job, and potentially threaten employment? Chopson believes that with these tools, architecture firms in the near future can do the same amount of work  in half the time. 

But, he suggests, with AI tools making it easier to figure out how to make projects pencil out, it’ll open up more opportunities to design projects. Combine that with the increased volume of work coming from disaster recovery, and he predicts more opportunities that may mean more employment and pay. Of course, it remains to be seen how the AI hype will play out, especially in the labor market. But as more projects like this proceed, it may begin to alter the expectations of clients.

“Sadly, too many developers are targeting to replace the architect, which I think is impossible,” said Greenberg. “The architect is the project manager, the one communicating with the builder; the homeowner; the city. We want to help them do the manual work they don’t enjoy doing so they can be creative.”

 Patrick Sisson is a freelance writer covering the AEC industry. He lives in Los Angeles.