Generative Artificial Intelligence and Building Design: Solving the Unsolvable

Building design has always required the collaboration of experienced designers and engineers who share ideas and bring together their skills and knowledge to create the best solutions for a particular project.

Now imagine collaboration that combines project data compiled from hundreds of designers across an organization — from recent arrivals to legendary retirees — with industry standards, local requirements and best practices seamlessly factored into the process to create dozens of viable options in a fraction of the time.

That is the impact that generative artificial intelligence (GenAI) is poised to make within the AEC industry. And it is coming soon.

The industry-altering impact of this technology is a game-changer. GenAI is being incorporated into new design tools that are able to use past project data to inform, compare, refine and recommend building design solutions. It is capable of quickly analyzing these large data sets from previous projects, while factoring in local codes and restrictions, and suggesting innovations to develop the foundation for a successful building design.

Until recently, GenAI felt more like a novelty than a tool that could move engineering forward. Early use cases demonstrated impressive capabilities — swiftly counting the number of private offices within a building, for example — but as a functional design tool, it was unconvincing.

Today’s technology is only getting better — and advances are happening quickly. Over the past year we’ve witnessed a huge jump in the design technology, evolving from narrow use cases to a useful real-world tool.

But it’s also important to remember that the technology will only ever be as good as the skilled engineers and designers who inform AI with their vast knowledge and experience.

The anticipated impacts of this technology justify a careful examination of where our teams are in the value chain, and how our business models align that value to each client’s desired outcomes.

As technology investor and Shadow Ventures CEO KP Reddy said, “Owners don’t buy drawings, they buy buildings.” As an industry, we need to consider how the process of creating buildings might change and, with that, our roles in that process.

In principle, AI is a tool that will make visionary architects, designers and engineers even more valuable in the design process, allowing them more opportunities to create buildings that maximize outcomes in sustainability, efficiency and overall aesthetic appeal.

One of the first concerns expressed about AI technology in any industry is that it risks removing human creativity from the process. But for AI to realize its potential in solving challenges that were previously out of reach, we need to focus this technology on solving new problems and not getting caught up in defending portions of our workflows that are automatable and commoditized.

By now most of us have read stories generated by tools like ChatGPT and will largely agree that while an impressive exercise, there’s something lacking in the final product — something human and personal.

The same is true with engineers and AI tools. AI will be useful for developing the backbone of a project design, blending known data about location, code requirements and purpose of a building to quickly create multiple preliminary designs. But the experienced professional’s insights, creativity and vision are needed to make it all work.

Rather than spending time determining the best locations for sprinkler pipes or where to place restroom facilities, engineers can use generated designs as a base from which to develop their own refinements and take projects to the next level.

Even before the process begins, GenAI can answer questions that were difficult to answer quickly and accurately, such as:

• How long would it take to build that design?
• How much carbon is embedded in that design?
• What are the tradeoffs in cost between these options?

It can take months of detailed engineering and design work to get to the point of discussing the pros and cons of building options A, B or C. Now, evaluating hundreds of structural models for a commercial building to identify the pros and cons of carbon, cost and constructability can be done at the concept stage.

With this head start, GenAI provides an opportunity to enhance the value of engineering and design services, while clients receive the answers they need to set expectations and determine next steps.

After the preliminary design stage is completed, GenAI tools ensure that the vision of the project remains intact throughout construction.

One construction monitoring tool can take pictures of a building under construction and then compare it to the engineering model. The tool’s precision is amazing — if a pipe fitting is off by even a few millimeters, AI can identify the variant and alert engineers about a problem that could cause construction delays downstream when components installed later won’t have sufficient clearance.

For example, a fire protection engineer assessing a project site will notice if the pipes are placed to code, but not necessarily see them in the context of lighting design that will be installed later. A poorly placed pipe could impact the lighting plan if it diverted from the design model. This tool can provide a holistic understanding of everything that could impact the design and pinpoint conflicts before they became headaches.

AI can also play a role in determining which materials would be most appropriate to use on a project. Perhaps there is material from a nearby build that may otherwise go to waste that can be repurposed at a reduced cost. Maybe more costly materials are available close to the project site that offer a lower carbon footprint — and thus a lower total cost to the project when considering code-required carbon offsets — than less expensive materials shipped from far afield.

There will be AI tools with the potential to uncover those opportunities with greater consistency, connecting those dots, saving money and resources, and incorporating more sustainable practices into construction.

If it takes less time to design, doesn’t that mean project fees will drop as well? Not necessarily.

True, some consultants will attempt to attract clients with lower costs — generating designs that are code compliant and functional, but derivative. At WSP, we believe clients will expect us to use this technology to solve the previously unsolvable.

First and foremost, clients invest in experience, expertise and confidence in the design, not on how much time is spent on creating the documentation.

GenAI is only as good as the data that informs it. WSP, for example, can bring its unique resources collected from thousands of projects around the globe to every design effort — information that will inform our unique approaches and create differentiated solutions from our competition. That is a better outcome both for our clients and for our industry versus a race to the bottom on cost.

It is why companies are wisely taking steps to ensure that their proprietary data is protected before diving into the AI world of design.

Being at the forefront of GenAI design doesn’t necessarily mean investing in the development of expensive cutting-edge tools. On the contrary, it could be unwise to purchase emerging tools, only to have them supplanted six months later by comparable mainstream toolsets from industry leaders like Autodesk, Bentley and Esri.

For WSP, collaborating with our ecosystem partners has been a successful model, letting their expertise guide our development. Hiring and retaining the best and brightest engineers in the industry remains WSP’s most important investment in AI.

And those best and brightest are currently exploring its potential. Field testing is taking place now across WSP to iron out technology issues and demonstrate the value of incorporating GenAI to clients.

Once the domain of highly specialized teams, GenAI is becoming democratized for broad use in building design. Future-focused engineering firms recognize that the day when AI becomes standard practice is near and embrace its opportunity to become the foundation of greater efficiency, productivity and inspired design.