“Building information modeling (BIM) may be another way firms can maximize the efficiency of their staff.” That was one takeaway from EC&M’s 2007 Top 50 Electrical Contractors Survey. Seventeen years later, adding artificial intelligence (AI) to BIM is shaping up to be another breakthrough way firms can maximize the efficiency of their staff.
“In the next 10 years, AI will transform the industry more than anything has in the past 100 years,” says James Barrett, vice president and chief innovation officer at New York City-based Turner Construction. “I think that applies to revolutionizing BIM, as well.”
EC&M’s 2024 Top 40 Electrical Design Firms Survey also indicates that this revolution is underway. When asked how they’re using AI, some respondents, such as Salt Lake City-based Spectrum Engineers, indicated they are using it to generate, analyze, and optimize electrical designs/BIM (e.g., conduit groupings, raceway sizing, routing, etc.).
“The industry has been using genetic algorithms in geometric design of building façades and general architectural designs for the past few years,” says Steve Germano, software development team lead at Rock Island, Ill.-based IMEG. “There are platforms like Testfit — targeted at building owners and architects — that are using AI and machine learning (ML) to iterate through permutations of leasable tenant space designs on lots for multifamily dwellings, commercial office spaces, etc. Conoa is another great platform that uses various ML/AI to help select products for buildings while analyzing total carbon footprint, taking into account product sourcing, proximity, manufacturing materials, and other factors.”
AI also could make it easier to use BIM tools by enabling “natural language” or “large language” interfaces. These let people type everyday terms to tell the AI tool to do something, such as: “Show all of the options for running conduit on this floor.” Another example is having the AI scour product databases to identify all of the luminaires that would work in a particular space based on requirements, such as features, size, cost, energy consumption, or price.
“There are a lot of resources that designers use on a daily basis that are in disparate locations on our intranet: a designer interface, code books, client requirements, project requirements, materials specifications, things like that,” says Adam Roth, BIM/VDC director at Lenexa, Kan.-based Henderson Engineers. “The benefit of having AI would be that you would just be able to ask questions and cut down on the transition between platforms.”
In a July 2023 EC&M article, New York City-based CannonDesign explained how it’s using Midjourney AI to combine images created with Revit and from other sources.
“[Suppose you] need a hospital lobby that’s warm, bright colors, double height with connecting stair, and you’re searching for that kind of thing on Pinterest or Google, and you’re not finding that image,” said Coffield King, senior lighting designer. “That’s where you can use the text and just tell it exactly what you want. It does that really well.”
So well, in fact, that some people consider AI on par with the mouse when it comes to user interface breakthroughs.
“Large language models are quickly becoming the easiest user interface humanity has had to computing power,” says IMEG’s Germano. “The last big change in computer interfaces was the mouse. Using natural language, the average person can leverage far more computing power than they could in the past.”
Work smarter, not harder
IMEG is so bullish on these benefits that it developed its own AI-powered chatbot named Meg, which currently answers more than 1,000 questions a day. In a recent podcast series, Germano explained how Meg makes it quicker and easier for employees to get information from a wide variety of internal databases that have been carefully curated and verified. That’s key because with any type of AI, quality output requires quality input.
“In the future, our engineers will be able to ask Meg, ‘Give me the total solar load on the west wing of the building,’” Germano says. “Upon receiving the answer, the engineer then will be able to instruct the AI assistant to increase the size of the air handlers to accommodate the stated solar load.”
AI also can improve productivity by freeing employees to focus on BIM tasks that can’t be automated.
“I envision AI being used to cut down on highly repetitive tasks and to help identify solutions that meet a prescribed/defined list of evaluation criteria,” says Matt Goss, senior vice president and MEP + Energy practice leader at Latham, N.Y.-based CDM Smith. “AI will most certainly increase design efficiency and allow engineers and designers to focus on more complicated tasks and solutions.”
Some people envision AI evolving into a personal assistant.
“I believe what’s going to happen is the tool almost being an AI assistant that will sit alongside the user and guide them through codes or things that are not readily available during the design process,” says Austin Stone, director of design efficiency and analytics at West Chester, Pa.-based Core States Group. “There’s a ton of them. It’s time consuming. It’s challenging in the sense of, ‘Where do I get this information?’”
In the future, AI assistants would ask and answer those types of questions based on parameters, such as a project’s location and what the authority having jurisdiction (AHJ) requires in addition to the NEC. This could be done automatically in the background and in real time, alerting designers when something could or should be changed — similar to how Microsoft Word can check for grammar and spelling using specific dictionaries and style guides.
“Instead of stopping and doing research, it’s going to be right there,” Stone says.
AI assistants also could use ML to understand its user’s personal work style. Those insights could enable the assistant to make suggestions without being asked.
“Having an AI to riff with is, in my opinion, what we will have readily available across the industry within the next few years,” says IMEG’s Germano. “Each person can have their own AI assistant with whom they can spitball ideas in real time and offload trivial tasks, so they can focus their time on more important and critical decisions.”
AI assistants eventually could be the primary way that people access BIM tools.
“I think, ultimately, our personalized AI are going to be the place we go to first,” says Turner’s Barrett. “All the solutions will be behind that in the sense that you’re not going to open up Revit or Navisworks. It’s not going to be about the solution. It’s going to be about what you are trying to get done: ‘Build me a design. Give me a cost. Give me a schedule.’ I think that’s the world where AI is bringing us. It’s not here yet — but it’s coming fast.”
All together now
But a lot of groundwork remains before the AI-BIM mash up can live up to its potential. Some of this will be internal work done by electrical design firms.
“Henderson is working on organizing and structuring our data, and mapping the different systems together: ERP, accounting, project database, BIM database, design database,” Roth says. “Mapping that all together is really where AI can transcend all those platforms and be able to give you analytics and insights that you would not otherwise have. We’re [also] focusing on structuring our data and getting it organized. AI is very, very powerful, but only in regard to structured data sets. If you have a bunch of unstructured data, then AI will not help you sort through that.”
On the vendor side, the groundwork can be grouped into two types. Electrical equipment manufacturers will need to put their product catalogs and other information into formats that AI-powered BIM tools can use. They’ll also need to create application programming interfaces (APIs) so the tools can connect to those databases.
“I think you’re going to see a lot more vendors developing tools that can be easily bolted onto an application that already exists,” says Core State’s Stone. “You’re already seeing that through Revit API.”
Integration takes time and money, which helps explain why some vendors may want to charge for that access.
“I don’t think you’ll ever get out of the ‘paywall versus free API’ [conundrum],” Stone says. “That’s going to be tough unless there’s some kind of lobbying or some kind of campaign from construction and AI [firms].”
The second type of vendor-side work involves the AI and BIM software providers.
“The majority of the AI out there is not compatible with BIM directly,” Roth says. “The nature of the BIM file and proprietary file formats, usually with the compatible software, is prohibited from running AI or any generative algorithms directly on the data set. So the majority of the time, we’re trying to extract certain parts of the data to run generative algorithms or evolutionary solvers on that data. That’s where you can really have some impactful decisions.
“I think the reason that the file format proprietary is that’s just the nature of the industry for the past few decades. The industry is trending toward more open data formats and interoperability, but we’re just not there yet.”
As AI proves its worth, BIM vendors will begin building it into their tools.
“BIM authoring tools like Revit have not integrated too much AI directly into their packages yet, but there are many trade-specific apps that integrate into these platforms that leverage ML/AI for modeling processes,” says IMEG’s Germano. “There also is a duct design tool, and another general MEP modeling tool that not only helps to draw the geometry but also considers total cost of installation based on the current design. This allows the design engineers to see the impact (both in sustainability and cost) at the time of design (the most economical time to make changes).”
Some BIM vendors are developing their own AI tools. One recent example is Autodesk’s Project Bernini, which the company says includes an “experimental generative AI model that quickly generates superior 3D shapes from a variety of inputs, including a single 2D image, multiple images showing different views of an object, point clouds, voxels, and text.”
Built to last?
For electrical design firms, another challenge is keeping up with all of the new AI tools and their capabilities. The selection grows every day — not only from BIM vendors, but also from IT companies such as Microsoft, AI specialists such as OpenAI, and startups such as Augmenta, whose initial product is designed specifically for the electrical industry.
“We get bombarded with new solutions,” says Kris Wahl, Turner Construction innovation manager. “It’s hard to cut through the noise now because everybody its touting AI capabilities on their websites. Every solution is now claiming to be AI powered. That’s a challenge.”
Another challenge is trying to determine which vendors will still be in business a year or two for now and which tools will continue to be upgraded like any other piece of software.
“[Some] will release their first tool set, and then that’s it,” Roth says. “There’s no progression from a company standpoint. We want to commit to a company that we can have a good partnership with and we know where they’re going and that they will progress their tool set along with the industry.
“It’s very important that we have somebody that’s in it for the long haul — that’s just not hoping to do the bare minimum and only doing to it get acquired by a larger company. We’ve seen that business model over and over again.”
