Cut through the noise

Evaluating technology for a property management company was relatively straightforward not too long ago. New software promised to solve a specific problem, say, streamlining accounting or digitizing maintenance requests, and leaders could weigh the investment against a clearly defined operational need.

Artificial intelligence has complicated that equation.

Amanda Cox, CPM®

Today, nearly every technology provider claims AI capabilities, and new startups appear almost weekly, promising to do everything from automating workflows to predicting leasing trends. Even general-purpose platforms like ChatGPT, Claude, and Gemini continue to evolve rapidly.

For property managers, the abundance of options has created a new challenge: deciding which AI tools deserve a place in their business.

“It’s easy to feel like you’re getting behind because there’s so much out there,” says Amanda Cox, CPM^®, president & COO at Bellaire Multifamily Management.

But Cox has learned to resist that pressure, noting that the better question isn’t whether an organization is adopting AI quickly enough, but whether the technology solves an operational problem, creates measurable value, and increases productivity.

In a technology market defined by rapid change, the advantage doesn’t necessarily belong to the company with the most AI tools. It belongs to the company with the discipline to determine where AI creates value and where it doesn’t. “In property management, moving fast doesn’t mean adopting every new tool,” Cox says. “Our philosophy is ‘fast in learning and deliberate in scaling.’”

Where to start

Regardless of business size, organizations need to build discipline in evaluating new AI technologies.

Cat Allday

“What is the business outcome you’re trying to drive?” asks Cat Allday, vice president of product at AppFolio. “Is it reducing vacancy days? Your renewal rate? Portfolio growth?”

Organizations that begin with desired outcomes tend to evaluate technology differently from those that begin with software demonstrations. Cox’s team has adopted a similar philosophy, assessing operational challenges before discussing vendors or products. Where does friction exist inside the business? Where are employees spending disproportionate amounts of time? Which problems meaningfully affect residents, associates, or clients?

Cox’s team establishes benchmarks before a pilot begins. “That might be how many hours we would save, improved leasing conversion rates, or faster response times,” she says. “ROI is ultimately the deciding factor, but more broadly than the cost savings, is there efficiency? Does it empower our team?”

That framework helps explain why some areas of property management have emerged as particularly useful for AI. “Leasing workflows, resident communications, maintenance, and accounting are where we’re seeing the strongest returns today,” Allday says. “They combine high volumes of repetitive work with direct impact on revenue, cost, and customer experience.”

For all the “revolutionary” attention AI receives, sometimes the most effective implementations involve simplifying and removing work.

As the managing director of WatchPoint Commercial Real Estate, a smaller commercial firm, David Barrow, CPM^®, recognizes that time and technology budgets must be used carefully. Rather than beginning with what AI can do, he considers the consequences if it gets something wrong.

David Barrow, CPM®

“I look at it very subjectively as what tasks have low risk if there’s an error, and what tasks have higher returns through automation,” Barrow says.

That philosophy has produced dozens of small improvements rather than one sweeping transformation. For instance, he’s using AI to help with summarizing lengthy association documents that once required careful manual review, drafting contracts using tightly controlled prompts and reference materials, and extracting key dates, meeting requirements, and responsibilities from governing documents that might otherwise take hours to navigate.

One of Barrow’s favorite use cases grew out of a practical operational problem. Using automation software connected to an email inbox, his company created a workflow to monitor incoming maintenance messages for emergencies. If an email mentions something like an active roof leak or storm damage, the system summarizes the message and immediately sends a text alert to the appropriate team members.

Similarly for Bellaire, rather than having associates manually enter invoice information throughout the approval process, AI now handles much of the administrative work, allowing employees to focus on quality control and exceptions. The cumulative time savings, Cox says, have been significant because the process touches multiple people throughout the organization.

Learn the value of “no”

While AI experimentation has been the norm for years, less attention has been paid to an equally important skill: knowing when an experiment has run its course.

Cox estimates that her company has killed roughly eight out of every 10 AI pilots, a rate she’s okay with. An unsuccessful pilot can reveal how employees actually work, what residents value, or where automation creates more friction than efficiency. The important thing is learning those lessons before an organization commits to a full rollout, she says.

For instance, Bellaire explored using AI as the initial point of contact for leasing inquiries, an application that made sense in theory: Prospective residents would receive immediate responses, after-hours inquiries could be captured, and leasing teams would spend less time fielding repetitive questions.

In practice, lead conversions declined, and feedback made clear that prospects still wanted to talk to people. Cox’s team pulled back quickly, then found a narrower role for the technology. Today, associates have ample opportunity to answer incoming calls before AI steps in after repeated rings, allowing the company to capture a lead it might otherwise miss without sacrificing the human interaction that helps build rapport.

Barrow has reached similar conclusions through his own experimentation. His enthusiasm for AI is tempered by skepticism toward features that seem to exist primarily so a software company can say it has AI. If a new capability simply accomplishes something he could already do easily, the label doesn’t make it valuable.

As AI becomes commonplace across virtually every software platform, organizations will need to become increasingly comfortable saying no and distinguishing between novelty and actual value.

People first

The public conversation around AI often gravitates toward automation and job displacement, but property management leaders are finding that the more immediate change is in the nature of the work itself.

For Barrow, that’s part of AI’s appeal. “What I’m most excited about is not having people spend time on things that they don’t need to,” he says. “You’re paying somebody a salary, and you want them to be able to achieve the purpose of their job. If you have limited support, like with small companies, you have to figure out how to minimize the amount of time you’re spending on everything.”

AI can free property management professionals from administrative work, allowing them to focus more on relationships, strategy, and proactive problem-solving. In this way, tools can change staffing not by eliminating employees but by raising the value of what only people can do.

Similarly, Cox sees technology automating and accelerating administrative work while leaving judgment, empathy, accountability, coaching, and relationship-building firmly in human hands.

Allday adds an important management challenge: Giving employees time back isn’t enough. Leaders need to decide how that reclaimed capacity should be used, redesigning work around more strategic and relationship-driven responsibilities.

Establishing guardrails

As organizations become more comfortable experimenting with AI, they also need to address data privacy, employee training, approval processes, cybersecurity, resident expectations, and organizational culture.

Boundaries need to extend well beyond a generic warning not to share sensitive information like residents’ personally identifiable information and owner or financial data, Allday says. “You need to know where the data goes, how it’s retained, whether it’s used to train those models, and who can access it,” Allday says, adding that leaders must understand the terms of service for these tools.

When AI is used for appropriate business information, Barrow prefers paid accounts with stronger controls around data use. In general, though, he says, “If you wouldn’t leave it sitting on your desk in an open area, you probably don’t want to enter it into AI.”

Organizations also need to decide what AI is authorized to do, what requires review, and where human judgment remains nonnegotiable. Transparency matters, too: If a resident is interacting with a company’s AI tool, should that be disclosed to the resident, and is there an easy path to a staff member when the situation requires empathy or judgment?

Guardrails are just as important for checking the work AI produces. Barrow once asked an AI tool to compare contractor bids. The system produced a polished chart showing bidders and prices across different components. “Except it got the lowest bidder wrong,” Barrow says. “How did it get that wrong? It was clear that the number was higher than the other bid.”

Cox encountered a similar limitation while testing lease file auditing software. The platform returned hundreds of exceptions requiring manual review, ultimately consuming so much staff time that employees could have performed the original task just as efficiently themselves. “AI is not a substitute for industry experience or sound judgment,” Cox says. “Some tools promise complete automation but still require a lot of human oversight.”

Building adaptability

Cox expects the AI evolution to eventually “be embedded in nearly every core function of property management,” she says. “We’ll see AI as standard in leasing and resident communications, predictive maintenance, automated financial workflows, real-time operational insights, and more.”

Accordingly, choosing the “right” AI tool at any single moment becomes less important than developing an organization capable of adapting as the technology changes.

The best AI companies today already understand that distinction, Allday says. “Organizations that are seeing a lot of success with AI are treating this as a business transformation and not a tech deployment.”

That transformation starts at the top, with executives using the technology themselves. It also involves the employees who understand how work actually gets done. Just as importantly, Allday says, leaders need to recognize that implementation isn’t a moment when one system replaces another, and the work is finished. “During these early phases of a rollout like this, you have to recalibrate,” she says. “It’s not like you just flip a switch and the AI takes over.”

The goal, ultimately, is to build what Allday describes as AI fluency: enough technical familiarity, operational understanding, and organizational confidence that a company can absorb technological change without scrambling to catch up.

For Cox, the same idea takes the form of an AI-ready culture in which employees understand not only how to use the technology but also why they’re using it and how it improves their work.

“I’ve relinquished the pressure of feeling like we’re falling behind and really focused on creating a disciplined framework that allows us to innovate responsibly,” Cox says. “Keep the focus on people, not the technology. AI adoption will succeed when your associates understand how it makes their job better.”