Zillow has put an AI assistant, built by EliseAI, into multifamily rental listings. HousingWire reported it as a chat tool that answers renter questions around the clock and books tours straight into a leasing team's calendar. That is useful. It also means a prospect's first conversation about your building may now happen with someone else's software, on someone else's platform.
That is a small example of something bigger. AI in multifamily property management is moving fast, and the owners paying attention are asking the right question: is this making my asset worth more or less?
I use AI every day in our work, so I am not here to argue against it. But I have also seen what it does when it is deployed without thought. The gap between those two outcomes is where this article lives.
The Pressure to Adopt, and the Trap in It
Every vendor, conference and trade article is telling operators to add AI. The fear of falling behind is real, and it pushes people to buy tools before they know what problem the tool solves.
Adopting a tool and having a strategy are different things. A strategy starts with a question: how does this protect or grow NOI, retention or asset value? If a tool cannot answer that, it is a cost, not an improvement.
When we look at how AI shows up in management companies, it tends to land in one of three buckets.
- Bolted on. AI added because everyone else did, with no clear goal. Some of it is harmless. Some of it removes the human contact that drives renewals.
- Used on purpose. AI aimed at specific, measurable jobs like monitoring equipment, drafting reports or sorting leads, with people reviewing the output.
- Built on expertise. A team with real financial and operating depth uses AI to cover more ground, faster, and knows when to ignore it.
The third is the goal. It is not about having the newest software. It is about knowing the work well enough to use the software well.
AI multiplies whatever you bring to it. Bring deep knowledge and you get faster, sharper work. Bring none and you get confident-sounding mistakes at scale.
Where AI in Multifamily Property Management Backfires
Chatbots on your most expensive leads
Imagine a busy weekend where your marketing generates a lot of high-intent leads. By Monday, a good number have already toured somewhere else. Your leasing team arrives to a queue of chat transcripts and very few booked tours.
The logic for chatbots is sound on the surface: prospects want an instant answer, and a bot never sleeps. Speed does matter. The trouble is that not every inquiry is equal. A person actively comparing three buildings and asking real questions is worth a lot of marketing dollars, and they want rapport and a straight answer.
Bots are fine at "can I have a dog" and "is parking included." They struggle with judgment calls: "I have two cats and one is an emotional support animal, will that work?" or "Your site says $2,200 but another listing says $2,150, which is right?" A stalled chat or a wrong answer can lose the lead on the spot.
There is also a control issue. When an AI layer sits inside a listing platform, the platform shapes the conversation, not you. You are paying for the lead but you may not own the first impression. The practical answer is not to drop listing sites. It is to build direct channels, your own site, your own brand, your own follow-up, so you are not dependent on someone else's assistant.
Our own approach is to let AI help with after-hours questions and routine inquiries, and to make sure a person picks up during business hours when someone is ready to talk.
Here is simple math to see why that matters. Say your building spends $30,000 a year on lead generation and a clumsy bot cuts the number of leads that turn into tours by a third. You have effectively thrown away $10,000 of that spend before the cost of the vacancy that follows. These numbers are an example, but the arithmetic is how to think about it.
Financial analysis with no one checking it
Owners are increasingly pasting their numbers into a general chatbot and asking for a plan. I have seen the results. The analysis looked polished and was wrong in several ways at once.
It suggested a rent increase far bigger than the market and existing leases could absorb, with no thought for renewal timing or what that would do to vacancy. It recommended cutting the management fee because it was "above average," without understanding the local labor market or what the fee covers. It showed an insurance saving based on a quote for a different coverage structure.
The tool did not malfunction. It did what these tools do, which is produce an answer that sounds authoritative. What was missing was a person who knew enough to spot the nonsense, and enough to ask the question properly in the first place.
Use AI for analysis if you like, but treat every number as a draft. If you cannot explain why an assumption is reasonable, you are not ready to act on it.
Residents who cannot reach a person
Routing all resident communication through automation is where the damage gets personal. Maintenance triage is the classic case. A resident reports water stains on a ceiling. An automated system reads it as cosmetic and schedules a painter. The actual cause is a roof leak, and by the time a person looks, the damage has spread.
The same applies to phone systems. A resident with a safety concern or a sensitive situation who is steered to "submit a form" will remember it. They do not blame the software. They blame the property, and they say so in reviews and at renewal.
Automation is fine at 2 a.m. for routine questions. For billing disputes, safety, complicated maintenance and anything emotional, residents need a person who can solve the problem. If you have not read our thoughts on this, see resident retention strategies for why the human touch pays for itself.
Where AI Earns Its Keep
None of this means avoid it. There are places where AI is a clear win.
Predictive maintenance
Sensors on HVAC units, water heaters and other major equipment can track performance against a baseline and flag a unit that is drifting before it fails. A technician then inspects and decides what to do.
That split is the whole point. The software watches thousands of data points no maintenance team could track by hand. The person makes the call on timing and priority. The payoff is replacing a small part on your schedule instead of an entire unit on a hot July weekend, plus fewer angry residents and less overtime. We cover the broader tool set in our guide to property management technology.
Pricing with a human in the loop
Revenue management systems weigh market trends, competitor pricing, leasing velocity and season to recommend rents and renewal offers. They are helpful, and the key word is recommend.
When our pricing tool suggests a number, a person checks it against what the software cannot see: new supply opening nearby, the condition of the specific unit, a resident who has paid on time for years and deserves some consideration. Sometimes the right call is to override. The companies that go wrong treat the output as gospel, push an increase on a resident who would have renewed happily, and end up paying for the turnover.
Good pricing is not about squeezing every transaction. It is about the best result across the whole building over time. The software gives a range. Experience picks the spot.
Lead sorting that sends the right people to people
Used well, AI can pre-screen inquiries, ask about timeline and budget, and send high-intent prospects to the leasing team quickly while keeping longer-term leads warm with automated follow-up. The key difference from the chatbot problem is the goal. This is a triage tool that gives your team better conversations, not a wall between your team and the lead.
Admin work that steals time from people
The most reliable gains come from automating work that is administrative: entering application data, building routine reports, scheduling, sending payment reminders, processing documents. If a leasing agent no longer spends half an hour typing the same details into several systems, that half hour goes to a follow-up call or a tour. If a property manager is not compiling reports by hand, that time goes to retention and planning.
This is also where the staffing question gets answered. Strategic use of AI does not cut the team. It moves their time to the work that actually moves NOI.
Use AI to take the paperwork off your people, not the people off your property.
Analysis speed, with judgment attached
In underwriting and due diligence, AI can run scenarios, scan operating statements and rent rolls for anomalies, and sort a long inspection report by urgency. That speeds up the first pass a great deal. It does not replace the person who decides whether an expense spike was a one-time roof job or a recurring problem, or whether an occupancy dip was a closed employer or a bad building.
The advantage goes to teams that can move fast and still be right, which takes both the tools and the experience to check them.
A Simple Test Before Any Tool Goes Live
Whether you are an owner approving a new tool or a manager proposing one, run it through four questions first. They take ten minutes and have saved us from a few shiny mistakes.
- What number is this supposed to move? Occupancy, renewals, days to lease, an expense line, hours of admin work. If nobody can name it, stop.
- What happens when it is wrong? Every tool is wrong sometimes. A wrong rent suggestion that a person reviews is a small problem. A wrong answer sent straight to a prospect or resident is a larger one.
- Who owns the relationship afterward? If the tool talks to your prospects or residents, make sure you can see the conversations, take over at any point and keep the contact information.
- How will we know in 90 days? Pick the measure and the baseline before launch. If the number did not move, change it or turn it off.
That last one matters more than it sounds. Plenty of tools survive for years on the strength of a demo and never have to prove themselves.
What AI in Multifamily Property Management Means for Your Data
There is one more thing owners tend to miss: where your information goes. Rent rolls, resident details, financials and lease documents are sensitive. When any of that is fed into a tool, you should know:
- Who can see it, and whether the vendor can use it to train their own products.
- Where it is stored, and for how long.
- Who is responsible if something is exposed.
These are fair questions for your manager and for any vendor they use. A good answer is specific. A vague one, or "it's all secure," is a reason to keep asking. And the same caution applies to you: pasting a resident's personal details into a public chatbot to save time is a privacy problem waiting to happen.
Where to Start If You Are Behind
If this all sounds like a lot, it does not have to be. The lowest-risk way to begin is with work that is repetitive, internal and easy to check.
- Reports and summaries. Let software draft the monthly narrative, then have a person correct it.
- Document sorting. Applications, invoices and inspection reports are good candidates because a person can verify the result.
- After-hours basics. Simple, routine questions answered overnight, with a clear handoff to a person in the morning.
- Monitoring. Equipment alerts and anomaly flags that point a technician somewhere, not replace one.
Leave pricing authority, resident disputes, screening decisions and anything legal for later, and keep a person in charge of them even then. You will learn more from three small, measured wins than from one sweeping rollout.
How to Tell Whether Your Manager Is Using AI Well
You do not need to understand the technology. You need to ask good questions and listen for specific answers.
Warning signs
- High-value prospects hit a bot first and cannot easily reach a person.
- Residents complain they cannot get a human during business hours for billing, lease or maintenance problems.
- The pitch is about cutting staff rather than improving results.
- No one can explain how outputs are checked. It is a black box.
- Resident messages feel generic and could apply to any property.
- The team lacks depth in the area where AI is being applied, like finance or marketing.
Good signs
- They can say where AI helps and where humans lead, and why.
- Human review is built into the process. Someone with experience signs off before action is taken.
- Prospect tours, renewal talks and complex resident issues stay human-led.
- They track real numbers: occupancy, renewals, expense ratios, resident satisfaction.
- They are open about limits and failure cases, and have a plan for when the AI is wrong.
Trade groups describe the same approach. In a letter to the Senate Commerce Committee, the National Apartment Association and National Multifamily Housing Council describe industry practice as including human oversight of decisions the AI recommends, with AI as one input among several. That is a fair standard to hold a manager to.
Seven questions to ask
- How are you using AI in leasing, and at what point does a prospect talk to a person?
- What experience sits behind each AI use? Who on your team understands that area deeply?
- How do you check AI output before acting on it?
- Which resident interactions are always handled by a person, and why?
- What has AI changed in occupancy, renewals or expenses, with real numbers?
- What happens when the AI gives a wrong answer to a prospect or resident?
- How do you protect the human relationships that support premium rents?
There are more along these lines in our list of questions to ask a property manager. The quality of the answers will tell you very quickly whether you are dealing with a thoughtful operator or a company that bought software because everyone else did.
Fair Housing Does Not Pause for Software
One more thing to keep in mind. If an AI tool helps screen applicants, writes your ads or answers prospects, the Fair Housing Act still applies to the results. A tool can produce an unfair outcome even when no one meant it to. You, and your manager, are still responsible. Keep a person accountable for decisions, ask vendors how their tools are tested, and have your attorney review anything that touches screening or advertising.
What Comes Next
The next wave is often called agentic AI: software that does more than answer questions and instead carries out a series of steps on its own, such as coordinating a move-in or reaching out to residents ahead of a renewal. The potential is real. So is the risk. In skilled hands it will take routine process work off the team. In unskilled hands it will repeat a bad process faster and at greater scale.
That is the point to hold on to. The need for expertise does not shrink as the tools improve. It grows, because the cost of a wrong assumption grows with the speed of the system.
What This Means for Your Property
If you own multifamily property, you do not need to become an AI expert. You need to hire for the fundamentals and ask sharp questions about the rest.
- Choose operators with real depth in finance, marketing and operations. Those fundamentals decide whether AI helps or hurts.
- Look for people who can explain where AI adds value and where it does not.
- Protect the human moments that drive retention and premium rents.
- Ask for evidence, not adjectives.
- Do not let hype do your buying.
At Next Level Property Management we use AI in analysis, market research, maintenance monitoring and reporting, and we keep people in leasing, renewals and anything where the relationship decides the result. We check what the tools tell us against what we know before we act on it. That is not a stance against technology. It is a plan to make the technology pay.
The companies that come out ahead will not be the ones with the most tools. They will be the ones that know exactly what each tool is for.
