AI in Property Management, From Vacancy to Renewal

AI in property management recommends rents, answers renters' routine questions, ranks applicants and points to the repairs a building will need next. At every stage of a tenancy the decisions with legal weight stay with the manager: the rent they accept, what is said to a renter and the applicant they turn down.

A Rent Recommendation Is Only as Good as Its Data

The first decision on an empty unit is its rent, and pricing software, often sold as revenue management, makes a recommendation for each unit from data on rents, leases and demand. In the products the US Department of Justice describes in its case against RealPage, the property manager chooses every morning whether to accept the recommendation, keep the previous day's rent or override it, and the same choice comes up for renewal leases. The complaint says those products let a manager accept many recommendations at once but override them only one by one, and that RealPage offered an auto-accept setting and pushed clients to use it.

The case itself is about that data. The DOJ and several states sued RealPage in 2024, alleging that its software combined nonpublic lease data from competing landlords to set recommended rents, and a proposed judgment published in December 2025 would keep that data out of live recommendations and limit model training to data at least 12 months old.

A manager weighing a pricing tool can ask two things of it: where its market data comes from, and whether overriding a recommendation takes as little effort as accepting one.

Answering Renters Before Anyone Visits

Once the rent is set, a leasing chatbot can answer the questions an office hears all day, such as whether the unit is still free, what it costs and when it can be seen, and book a viewing in the evening when nobody is at a desk. What the bot says is said on the landlord's behalf, and the Fair Housing Act makes it unlawful to make any statement about renting a home that indicates a preference based on race, color, religion, sex, handicap, familial status or national origin, and a bot's answer is such a statement.

A virtual tour lets renters look before they ask, showing the whole unit at their own pace, which helps someone moving from another city who can't drop by after work. The Harvard study of virtual tours looked only at homes sold, leaving rentals for later research, and what it found for buyers was screening: a tour helps people rule a home out, so they never book the visit.

Before a bot goes live, a manager wants to know when it hands a conversation to a person, and who reads what it told renters until then.

Tenant Screening Stays the Manager's Decision

The viewing produces an applicant, and the applicant a screening report. Some screening companies add a score or an accept or decline recommendation to the report. The recommendation doesn't move the decision: refusing to rent because of race, familial status or another ground the same Act protects is unlawful, however the refusal was reached.

When a manager turns an applicant down based even partly on a screening report, the Fair Credit Reporting Act requires a notice that names the screening company, says it didn't make the decision, and tells the applicant they can get a free copy of the report within 60 days and dispute what is in it. The manager sends that notice, so before acting on a score they need to see what it rests on.

Repairs, Energy and the Renewal

With a tenant in, the routine work is the easiest to hand over. Payment reminders, receipts and late notices follow rules the manager sets once, and a maintenance request can be sorted by urgency and routed to the right contractor without anyone retyping it.

Predicting repairs is harder. Predictive maintenance decides when to service equipment from its measured condition, where preventive maintenance follows a calendar, and software that predicts a failing boiler or air handler rests on the same idea. The Department of Energy's operations and maintenance guide, written for federal facilities in 2010, before today's AI tools, estimated that a working predictive program saves 8% to 12% over preventive maintenance alone. Of any tool that predicts failures, ask what condition data it needs from the building's equipment and what collecting it costs.

Energy is the other place software acts on its own. A smart thermostat learns when a home is empty and turns the heating or cooling down, and ENERGY STAR certifies models based on field data showing they save energy. Heating and cooling make up almost half of the average American household's energy bill, more than $900 a year.

At renewal the same pricing software recommends the rent again, with the same choice to accept, keep or override it, so the two questions from the first stage come back with it.

A Rental Listing Built From the Rooms Themselves

If the tenant leaves instead, the unit needs a listing again. Giraffe360 has no part in pricing, screening or repairs; its part is that listing, made from a visit to the empty unit with the Giraffe360 camera.

The virtual tour is shared from Tour links on the project, as a link for a listing or a message, a QR code, or an iframe on the manager's own website, so renters can look round before they enquire.

The floor plans come from the same capture. Setting Room measurements to Area labels each room with its size, and the area calculation standard can be set to the one the market uses, such as United States. Inside the tour, Show room area adds room sizes to the mini-map.

The listing text can start as an AI draft, which the manager checks just as they check a recommended rent or a screening score. The Property Description writes a draft from the rooms the manager keeps switched on, the total area and the places on its Location tab, in the writing style they choose. It writes in English only, and the product's own guidance is to read the draft for accuracy before it goes into a listing, since a generated draft can make claims about the location that need checking.

Renters who want to ask something can use the contact form in the tour, turned on under Info and Contact form. Their enquiries arrive in Leads alongside those from protected tours, gated tours and property websites, where the manager can filter them by project, source and date and download the list as a .csv file.