Artificial intelligence is playing a bigger role than ever in facility management and building operations, from finding opportunities to optimize building systems to simplifying decision-making. It can empower building owners and operators to work smarter—but it’s not a silver bullet, warned Maureen Ehrenberg, president, commercial division, of Lessen, whose Digital FM+ tool uses AI to help create maintenance plans and improve building resilience.
“Most people believe it’s a ubiquitous solution and it has all the right information to convert inputs and outputs into outcomes. It doesn’t,” Ehrenberg said. “It has to be trained to deliver an outcome. You can’t skip that step. You’ve got to make sure your model’s right and it’s learning.”
The right AI investment can save time and money—but a poorly engineered solution can cause more issues than it solves. Here’s what you need to know about using AI in facility management.
Where to Use AI in Buildings
Today’s AI models are fairly good at a few basic tasks, including:
1. Reviewing and analyzing large data sets of different formats. “It’s very good at analyzing tremendous amounts of data,” explained J. Wickham Zimmerman, CEO and cofounder of OTL, a provider and preserver of habitats, waterscapes, landscapes, and natural environments. OTL designed an interactive water feature at EpicCentral in Grand Prairie, Texas, that uses AI to respond to visitors’ movements with water jets and changing light colors. “With the use of cameras, it can identify and locate objects and movement and then translates that data into output signals that control any equipment running DMX, which nothing else at this point can do effectively.”
2. Streamlining repetitive processes. AI can easily input invoice information or delivery receipts, Zimmerman said. Facility managers can use AI to handle mundane data entry tasks and free up time for more complex issues that require a human to solve. AI also excels at analyzing large data sets much faster and it is able to identify patterns not evident to humans. For example, AI can be used to analyze energy usage over time to identify potential sources for improving efficiency or saving money.
3. Retrieving historical asset information. Some AI tools can retrieve equipment information from the asset registry of your computerized maintenance management system (CMMS) and assist with troubleshooting, Ehrenberg said. “If you’ve got that registry of equipment, it can help you in the way you set up your analytics. You can easily pull training data and visuals of it,” Ehrenberg said. “AI is great at taking the data you’ve got in the system and connecting the dots.”
This capability can also help with work orders, Ehrenberg added. If a user having an air conditioning problem tells an AI tool which building they’re in, the tool could auto-complete some of the information that will help the facilities department fulfill the request, such as the tonnage of HVAC equipment. “You can remove that task of someone having to complete all that information,” Ehrenberg said. “But it’s only when you have that information in the AI that it can apply the solution with a sense of certainty.”
Where Not to Use AI
AI tools have come a long way, but they’re not perfect. One key consideration to remember is that AI tools need to be trained in depth, Ehrenberg said.
“If the model’s not trained, you can get really bizarre results. It really needs to know ‘This goes with this’ to connect those dots,” Ehrenberg added. “It’s continuously teaching itself and learning.”
This can be a challenge in buildings, which can vary significantly from each other. Spaces that are very similar to each other—say, a branded retail design—are easy to learn from, but comparing one office tower to another can be a different issue entirely. “When you look at a typical office building and the tenant spaces in them, the years they’re built, and the materials they used—buildings are unique,” Ehrenberg said. “The equipment, assets, and designs are unique, and where they operate is not consistent, so the weather is different, the climate is different, and the way the climate impacts the equipment is different. Luckily, you can tap into that data and bring it in, but those dots have to be connected, or you can get a false result out of the AI. Those pieces need to be considered.”
Generative AI—a type of AI that uses its learning to produce new text, images, videos, or other content—also has limitations, Zimmerman said. There has been much talk of using generative AI to assist with building design and construction, but it’s not yet at the point where it can assist an architect with a unique design. “AI is very good at being able to detect patterns and make predictions based on past information, but starting from scratch on design projects, it’s not there yet,” Zimmerman said.
Make Smart AI Investment Decisions
Building owners and facility teams looking to invest in an AI solution should proceed carefully, Ehrenberg advised. “Take a step back and think about your operating model. What are you trying to solve for?” she said. “There are some great technologies out there, but they might not be something you would consider particularly useful, or the cost is not going to have the business case for you to implement.”
Once you understand why you’re looking to AI and what problems you want it to solve, it’s critical to understand what data the AI tool will train on and determine whether the data you have is sufficient.
“To get efficiency out of any kind of technology, first you have to have good data to start with, and it’s the same thing with AI,” Zimmerman said. “The old adage of ‘Garbage in, garbage out’ is more true today than probably ever before.”
If you’re missing something—for example, climate data that could help draw a clearer picture of how your equipment is affected by weather—consider how you can source what you need, Ehrenberg said. “If I’m missing a piece of two, can I buy that and integrate it? Can I subscribe to something and get that data? Before I deploy AI, should I change the processes that I’m deploying on site? Maybe I’m not collecting enough of the data I’m going to need for the AI, and I’m going to have to change my internal processes first because I have too much missing information.”
Facility and property professionals preparing to invest should also ensure they’re getting a thorough demonstration of any new AI tool from the vendor, Zimmerman added. Ask to use the tool yourself to see whether it will work for you. “Try to actually use it and understand what it is really going to do, because sometimes, what the salesperson may tell you may not be your actual experience with it,” he advised.
AI technology is changing at a rapid pace and stands to improve by leaps and bounds over its current state. It’s important not to jump in without having fully vetted the AI tool you’re thinking of investing in, but at the same time, it’s crucial that you don’t write off AI tools completely simply because they incorporate AI technology.
“I think we should see efficiency improve, not only on the personnel side, but on energy consumption, resolution of maintenance issues, and predicting ‘We know we’ve got this piece of equipment in use, and it’s got this life based on what the manufacturer says. Based on the way we’ve used this piece of equipment in the past six years, in three months the bearings in this motor are going to have to be replaced,’” Zimmerman predicted. “Those types of things, it’s going to be very good at, and we’ll become so accustomed to it that we’ll wonder how we did without it. I think we’ll be at that point in no more than five years, maybe even less.”
