How AI Is Making Multifamily Maintenance More Efficient
Property managers are expanding its role while keeping it on a short leash.
As commercial real estate companies expand the testing and adoption of artificial intelligence applications across operations, they are anticipating the day that AI leaps from automating mundane tasks and suggesting simple solutions to making more complex decisions.
In the apartment industry, the prospect of that next frontier is intriguing to property managers who want to boost maintenance efficiency as competition for a smaller pool of skilled technicians accelerates. While property managers are embracing current AI capabilities that can help technicians troubleshoot problems, they want to further support hands-on maintenance staff with more robust applications, from delivering detailed instructions to technicians to performing equipment inspections.
But property managers want to first prove that the technology can make those more autonomous decisions before giving AI more independence.
“AI can process information and provide us with good information and recommendations in an efficient manner that we could not do otherwise,” said Ken Murai, founder and CEO of FacilGo, a provider of AI-enhanced solutions for rental property owners and managers. “And while it is making decisions today, they are small decisions. I don’t think we’re in a place right now where we can allow AI to make decisions for us.”
READ ALSO: How Predictive Maintenance Is Reducing Risk in Multifamily Operations
Indeed, Watts Realty, an apartment and commercial property management firm based in Birmingham, Ala., is encouraging its technicians to use Claude to instantaneously identify equipment brands, models, serial numbers and error codes that can help personnel discover solutions and make speedier repairs, said W.A. “Chip” Watts IV, president of the firm. That’s a big improvement over a technician spending 30 minutes locating and then thumbing or clicking through a manual, he added.
But human confirmation of AI performance remains essential, he asserted, especially when a maintenance issue presents an acute safety risk, such as the loss of air conditioning on a 100-degree day. “If AI doesn’t take that into account, then the problem may not be flagged as urgent when it truly is,” Watts said. “So we have to have human oversight because I’m not comfortable at this point in time to let AI go and do everything on its own.”
More with less
Measured by job postings, demand for maintenance supervisors and technicians in the second quarter of 2026 grew 12.7 percent and nearly 9 percent, respectively, over the previous year, according to the National Apartment Association’s latest labor market report. The maintenance personnel staffing challenge is being exacerbated by the fact that new technicians lack the skill sets and experience to work on equipment that is increasingly embedded with new technologies, said Mindy Gronbeck, president of the Institute of Real Estate Management.
In some cases, that is forcing more property managers to outsource repairs, added Gronbeck, who is based in Boise, Idaho. “Everyone in our industry is having difficulty retaining skilled laborers and are trying to do more with less,” she said. “In Boise, maintenance people who have the skill set to do the jobs are being poached from multiple properties.”
In addition to helping technicians quickly troubleshoot equipment malfunctions, AI chatbots currently have the capacity to log maintenance issues and make recommendations to resolve or prioritize them and execute work orders, observers say. At CAPREIT, AI chatbots responding to reports of a leak can ask questions about the depth of the standing water, compare the severity of the problem to a set of guidelines and inform technicians to check lower floors if a leak is reported on an upper floor, said Savas Karas, chief technology and transformation officer for the real estate company, which is based in North Bethesda, Md.
“Ninety percent of calls are what I would call routine – they’re predictable,” Karas added. “And if they are routine and predictable, you can create a case and give AI guidelines and boundaries that it should work within.”
Keeping tabs on AI
How well AI performs even in relatively straightforward automations still depends on the data it’s crunching, which not only includes the amount of information and the quality of the information. Random, irrelevant or inaccurate data can become “noise” that interferes with AI’s education. But property managers acknowledge that the technology can make mistakes even in clean data environments.
“The key for AI is, has it had enough positive outcomes to illustrate that it understands the process, do you have confidence that it’s doing the right thing and is it doing the right thing repeatedly?” Karas said. “If so, then you can be more comfortable counting on the outcomes. But AI still does unusual and odd things at times, so you still need a human in the loop to validate it.”
Among other risks, anytime technicians are relying on AI when working with residents, errant recommendations or work orders that fail to address a problem could lead to perceived bias and run afoul of Fair Housing Act standards, Murai said.
He recalled a recent story about a window covering in need of repair being ignored because the technician didn’t like the resident. AI could reduce that type of subjectivity, Murai said. “But knowing AI can also run off the rails at times,” he added, “if you don’t put barriers in place, you could run into virtually the same situation.”
