THE PROFESSION OF ARCHITECTURE is at a crossroads. According to Carl Christensen, co-founder and CTO of Spacemaker AI, the whole building industry, while digitized, lacks critical building blocks that would enable the AEC industry to make big leaps.
Describing his initial reactions to how architects deliver value in early-stage development for clients, he says, “I didn’t really see the building blocks. People worked in digital tools, but some stakeholders work in drawing applications while others work in Excel.”
New to the AEC industry and with a background as a computer scientist, Christensen used the word “shocked” when he met Spacemaker CEO and co-founder Håvard Haukeland. When Haukeland, an architect, explained how architects deliver early phase work for clients, Christensen assumed Haukeland was doing something wrong.
Driven by Experience
“I was basically working [as an architect] with the kinds of projects that users today are using Spacemaker AI to solve,” says Haukeland. “I was really frustrated that I was not able to deliver what the clients really wanted, and I started discussing this with Anders, who I had known before.”
Anders Kvale is the third co-founder with a business background who saw Haukland’s problems—which involved working with clients who have either acquired a building site or evaluating one—as an interesting problem to look into.
This is a global problem; we need to build so much more in every city. When we build more densely, it is technically more difficult. There are more stakeholders and with them comes complexity.
The core challenge was how to optimize the development potential of building sites by breaking free of the traditional toolsets and processes that were leaving Haukeland disappointed. Kvale introduced Haukeland to Christensen, who has deep experience in software development and the digitization of industries.
“So the problem definition was really very clear,” says Christensen. “However, the more you looked into it, the more the complexity grew.”
A Worthy Global Goal
This complexity seemed ideally suited for AI and machine learning, but larger than that, the timing to help urban and architectural development make a big leap was now. “This is a global problem; we need to build so much more in every city,” says Haukeland, referring to massive urbanization change taking place around the world. “When we build more densely, it is technically more difficult. There are more stakeholders and with them comes more complexity.” (Image 01)
But the development process today needs re-engineering itself, and the Nordic team knew this. “So architects have their knowledge expertise, and so too do engineers, and the developer is focused on their goals, but none of them are really communicating in any kind of shared space where they can understand each other in any meaningful way,” says Christensen. “So we figured we needed to bring all this information together in a way that is transparent to the other stakeholders.”
With a clarified vision of the problem Spacemaker AI was aiming to solve, the company began developing its namesake solution and directing it to architects. But that is not where they found success. Architects’ usage of information technology at a budget level is less than 1.5 percent of revenues in some cases, so Spacemaker AI found greater success going directly to large property development firms, who passed Spacemaker technology on to their internal designers and architects, and external consultants.
Value in the Early Phase
Spacemaker AI describes itself as “early-stage planning reimagined.” The cloud-based solution centralizes the various knowledge and issues typically siloed by the A, E, C, and O professionals in the AECO world. It empowers stakeholders to optimize the development potential of building sites. And it uses artificial intelligence (AI) and machine learning to achieve this.
Haukland told me on our call that a McKinsey report says that only 20 percent of technology investment in the building industry is directed at early phases. Still, he notes another report that finds that it is the early phases where 50 percent of the value creation happens.
You mentioned the revolution of hand drawing to CAD, from CAD to BIM, from BIM to Cloud. And now we have ‘outcome-driven’ design. And that’s now the fourth revolution.
The problem is that value creation can be higher and achieved at faster rates if machine intelligence can augment the heuristic processes human designers employ. Spacemaker AI’s goal isn’t to replace human architects, planners, and urban designers but rather augment them with “over the shoulder AI,” as Haukeland and Christensen call it.
Christensen says of today’s workflows for this early-stage work, “there is a lot of manual labor inside a digital tool, but there is no added value across projects, everything is bespoke.” This takes us back to those building blocks he says the industry is missing. He means that without digital connections between data from various stakeholders and the meaning and values of that data from each stakeholders’ perspective, there is no value-chain present, no value creation. “It’s really hard to extract value and knowledge from such a disconnected process,” he says. “If you are trying to analyze processes and systems, that is just shockingly far from what most industries have already accomplished.”
The Fourth Revolution
Indeed, the AECO industry is profoundly behind taking true advantage of the ICT (Information and Communications Technology) revolution. (see: Architosh, “Perspectives on BEST of SHOW 2018: Perez’s Model of Technological Revolutions—BIM, CDEs, and VR,” 12 July 2018) But the current constellation of digital technologies, particularly when paired with AI and machine learning, may finally offer a step-change in productivity gains.
“You mentioned the revolution of hand drawing to CAD, from CAD to BIM, from BIM to Cloud. And now we have ‘outcome-driven’ design, says Haukeland. “And that’s now the fourth revolution.
Spacemaker AI helps users arrive at optimal outcomes for given building sites, tuning development projects to particular sets of criteria in optimized formations that will yield better outcomes. (Images 01 – 02)
The software does this by both generating and evaluating designs against inputted criteria. The system can analyze solutions for sunlight, daylight, views, wind, traffic noise, energy efficiency, and all the zoning setbacks and FAR (floor-area-ratios) impacting projects. (see Image 03 for Wind)
Spacemaker AI can optimize designs for over 100 criteria types, including doing all the boring calculation work like tenant lease area and building area calculations. The user inputs the criteria but also determines which criteria to emphasize. You let Spacemaker AI help solve for quality outdoor spaces, for example, by emphasizing sunlight utilization and controlling shadows. It can solve for microclimates and views to the river, for example, and at various combinations of importance.
It’s really hard to extract value and knowledge from such a disconnected process. If you are trying to analyze processes and systems, that is just shockingly far from what most industries have already accomplished.
As for the AI, Christensen says, “it is incrementally supportive of what you want to do but not taking over the wheel,” making a driving analogy. So, in other words, in the same manner, that a Tesla can provide some autonomous driving functions with human oversight, Spacemaker uses artificial intelligence to augment human designers but requires human oversight at multiple levels.
The software is flexible in terms of how it can relate to existing urban design and architectural workflows. “Users can import previous designs in SketchUp or Revit format and run analyses on them using the systems’ AI-based engine for testing scenarios,” he notes. “And once completed, users can export their project out to more precise design programs like SketchUp or Revit.”
Designed to be interactive as part of the overall process, Spacemaker AI enables the designer to generate and test in multiple iterative steps, fine-tuning both criteria and possible design options. The system isn’t there to find the ultimate solution for you by itself.
Christensen says there is confusion in the industry about solutions like Spacemaker AI. He says that “black box AI seems very attractive at the onset,” but that is not what people really want. “You cannot iterate on it,” he says, referring to the way the designer can take what Spacemaker AI has produced and understand why it got there.
The Autodesk Acquisition
Last year Autodesk acquired the Norway-based company. Over the past few years, the US design software giant largely focused on construction professionals within AEC—with a series of stunning acquisitions. In the area of new software for designers and architects, the Spacemaker AI buy was seen as a positive move.
“We never talked about an acquisition,” says Haukeland. We always planned for a standalone journey.” He says the company, at 115 people as of the time of our conversation, has been solely focused on its mission. “We are driven by that story. Everyone on the team is driven by how we can optimize the process.”
With that focus, Autodesk is leaving the company to operate very autonomously. The company stays in Norway, though it has staff in remote locations globally.
“Right now,” Haukeland adds, “the obvious synergy with Autodesk is on the mission of being part of that fourth revolution and of making a digital tool that everyone uses.”