The Honest Questions to Ask Before You Buy Any Leasing Software

Before you buy any leasing software, the questions that matter most aren't about features — they're about responsibility. Who owns it when the AI is wrong? Is it fair-housing safe? How does it stop scammers from grabbing a lockbox code? When does a human take over? How fast can you leave? And what proof exists for a portfolio your size?

Why do most leasing-software buying decisions go wrong?

The demo shows you the happy path. A lead comes in, the AI responds instantly, a showing gets booked, everyone smiles. What the demo never covers: what happens when the AI answers a question incorrectly, when a setting changes without warning, when a scammer clears your "verification" step and walks into your vacant unit, or when the vendor gets acquired and your price doubles.

Buyers evaluate on feature lists. Failures come from responsibility gaps, edge cases, and exit terms — none of which appear on a feature grid.

Across more than 112 discovery conversations with residential property managers, the same pattern surfaces: the complaints that eventually drove a switch were almost always rooted in things nobody asked about at purchase. A tool used for a decade described as "slow, laggy, clunky" — not a new problem, but one that calcified unnoticed. Settings that "just mysteriously change" after updates. An ID requirement "added without warning" mid-application, killing a qualified prospect. These aren't fringe cases. They're the things a good question catches before you sign.

The seven questions below map directly to that failure pattern. Bring them to every demo. A vendor who answers them confidently is worth more of your time. A vendor who deflects is telling you something important.

Is the vendor's AI actually fair-housing safe — or are you taking on the liability?

The short answer: you may be taking on more liability than you think, regardless of what the vendor claims.

In May 2024, HUD issued guidance confirming that the Fair Housing Act applies to AI and automated tools across the entire screening and leasing process — including showing scheduling, inquiry responses, and tenant communication. Critically, both the housing provider and the technology vendor are responsible for violations, including disparate-impact discrimination that produces discriminatory outcomes with no discriminatory intent.

"Our AI is unbiased" is not a compliance answer. It's a talking point.

A good answer looks like this: the vendor can walk you through exactly how the system avoids steering (directing prospects toward or away from specific units based on inferred characteristics) and disparate-impact outcomes. They're explicit about shared compliance responsibility — not vague about it. They've tested for it.

Question to bring to your demo: "Show me how your AI avoids steering and disparate impact — and who carries the compliance responsibility when something goes wrong?"

What happens when the AI is wrong — who's on the hook for what it says?

This question has a legal answer now, not just a theoretical one.

In 2024, a Canadian tribunal ruled in Moffatt v. Air Canada, 2024 BCCRT 149 that a company was fully liable for incorrect information its chatbot gave a customer — rejecting the argument that the chatbot was a "separate legal entity" with its own responsibility. The company owned what the bot said.

The implication for property managers is direct: if a leasing AI quotes a prospect the wrong rent, misstates a pet policy, or misrepresents availability, you own that representation. The vendor doesn't.

And AI gets things wrong more often than most demos suggest. The Vectara Hallucination Leaderboard (continuously updated) benchmarks the top large language models on grounded summarization — a constrained, document-anchored task that represents the best case for factual accuracy. Even under those ideal conditions, hallucination rates range from about 1–5% across leading models. On open-ended queries, where the model has no source document to stay faithful to, rates climb substantially higher. "We use AI" is not a guardrail. It's a description of the risk.

A good answer: the AI answers only from your listing data and approved rules. When it's uncertain, it says so — and routes to a human rather than guessing. Every answer it gives is logged and auditable.

Question to bring to your demo: "Am I legally bound by what your AI tells a prospect — and can I pull an audit log of every answer it gave?"

[[cta]]

How does the tool stop scammers from getting a lockbox code?

Rental scams are not a minor inconvenience. A December 2025 FTC Data Spotlight found nearly 65,000 rental-scam reports and approximately $65 million in losses since 2020, with a median loss of $1,000 per incident. Adults aged 18–29 were three times more likely to report a rental-scam loss than older adults.

Scammers don't just defraud prospective tenants. They target the lockbox code itself — gaining access to vacant units, which then become occupied by squatters or get stripped.

One operator managing roughly 1,100 units described the decision plainly: they turned off self-showings entirely after "nothing but squatters, nothing but scam artists grabbing codes." The problem wasn't the self-showing format. It was that ID verification came after the code, or not at all.

A good answer: a real identity step — government-issued ID or bank-level verification — clears before any access code is issued. Not optional. Not after the fact. Before.

Question to bring to your demo: "What ID step does a prospect clear before a lockbox code is issued — and is that step skippable?"

When does a real human take over from the AI?

Automation that books showings at 2am is valuable. Automation that doesn't know when it's out of its depth will cost you deals.

The case for fast AI response is well-documented: the MIT/InsideSales Lead Response Management study (15,000 leads across 100,000 calls) found that contacting a web lead within five minutes makes qualification roughly 21 times more likely than waiting 30 minutes — a finding later popularized in the Harvard Business Review — and that lead quality drops approximately 80% after that five-minute window. Speed matters — but speed alone isn't the answer.

From those same 112 conversations: an operator who tested a call-center AI product described its agents as "generally having heavier accents… not actually that consistent." The frustration wasn't about accent — it was about inconsistency and the absence of a clean human fallback when the AI stumbled. Prospects dropped out. Showings didn't book.

A good answer: the system has clear, explicit escalation rules. Complex questions, edge cases, and high-intent prospects route to a named human contact, fast, with full conversation context handed over. Not "contact us," a name and a timeframe.

Question to bring to your demo: "What exactly triggers a handoff to a human — and how fast does that happen?"

Will it actually work with your PMS — or just "integrate" on a slide?

"Integrates with [your PMS]" on a slide is not a sync. Ask what specifically moves between systems, in which direction, and how often.

A one-way push of listing data from your PMS to the leasing tool is not the same as a two-way sync that writes applicant status, showing results, and lead disposition back into your system of record. The first is a data feed. The second is an integration. Only one of them keeps your records accurate without manual intervention.

The "settings just mysteriously change" complaint — a recurring theme in those 112 conversations — often traces back to a shallow sync that drifts. Your PMS updates availability; the leasing tool keeps showing the old listing. A prospect books a showing for a unit that's already leased. Now you're making a phone call you shouldn't have to make.

A good answer: named, documented two-way sync with your specific system — not a logo on a slide, not "via Zapier," not "on our roadmap." A live demonstration of what moves in both directions, with your PMS, today.

If you're evaluating integrations with specific platforms, see our deep-dives on AppFolio, Buildium, RentVine, and DoorLoop for what two-way sync actually looks like in each.

Question to bring to your demo: "Show me the live two-way sync with [my PMS] — not a logo on a slide. What writes back into my system?"

What's the real contract — and how fast can you leave if it doesn't work?

This is the question vendors most hope you'll forget to ask.

The showing-tool and leasing-software market has seen significant private-equity consolidation in recent years. Operators who've come through those transitions describe the same pattern: price increases after acquisition, feature development that stalls while the new owner extracts margin, and support that degrades. Several operators in those 112 conversations cited a PE-backed roll-up as the direct trigger for their search — they wanted out, and they wanted to understand what "out" actually meant before they signed their next contract.

It means three things: how long until you can leave, who owns your data when you do, and whether your pricing is locked in if the company changes hands.

A good answer on pricing uses model-based framing (per unit, per month) with the terms clearly stated in writing. A good answer on exit gives you a real opt-out window — not just a technical out after a multi-year lock — and confirms that your contact data, showing history, and applicant records are portable on exit.

Question to bring to your demo: "How fast can I leave, and what happens to my data and my pricing if you're acquired?"

What proof can the vendor show that this works for a portfolio like yours?

Logos are not proof. A wall of brand marks tells you the vendor has sold to someone. It tells you nothing about whether it worked, for whom, or at what scale.

Multifamily is not residential. A 500-unit Class A apartment community has different vacancy pressure, different prospect behavior, and different operational constraints than a 300-door scattered-site residential portfolio. A case study from the former does not predict results for the latter.

Ask for specific, named (with permission) outcomes from a comparable operator — your size, your property type, your market. Ask for the metrics: response time, showing conversion, vacancy reduction. And ask to speak with that reference directly, not just read a quote that the vendor wrote.

Question to bring to your demo: "Show me results for a residential portfolio my size — not a logo wall. Can I speak with that reference directly?"

The one-page question checklist to bring to every demo

Print this out. Ask every question. A vendor who answers all of them clearly has thought through the hard parts. A vendor who deflects on two or three is showing you something the demo was designed to hide.

For reference, here's what good answers look like in practice. LetHub, for instance, responds to rental inquiries in under 30 seconds, requires bank-level ID verification before any showing access code is issued, syncs with all major PMS platforms — AppFolio, Buildium, RentVine, DoorLoop, and others, has clear human-escalation routing built into the product, and operates on a per-unit monthly pricing model with a 30-day opt-out. Not every vendor will match all seven questions. That gap is the information you're buying with the checklist.

For more on what the best AI leasing tools have in common, see our category guide on what to look for.

[[cta2]]

Frequently asked questions

What questions should I ask a leasing software vendor before buying?

Ask about fair-housing compliance and who owns AI liability, what ID verification happens before a code is issued, when a human takes over from the AI, how the PMS sync actually works (one-way vs. two-way), what your exit terms are, and what proof exists for portfolios your size. Those seven questions surface the gaps that a standard demo is designed to skip.

Is AI leasing software fair-housing compliant?

Not automatically. HUD guidance from May 2024 confirms that the Fair Housing Act applies to AI tools across the entire leasing process, and that both the housing provider and the vendor share compliance responsibility — including for disparate-impact outcomes that occur without intent. Ask each vendor to demonstrate how their system handles steering and disparate impact before you commit.

Who is liable if a leasing AI gives a prospect wrong information?

You are. The 2024 ruling in Moffatt v. Air Canada (2024 BCCRT 149) established that a business is fully liable for what its chatbot tells a customer, rejecting the argument that the bot is a separate legal entity. If your leasing AI misstates a rent, pet policy, or availability, that misrepresentation is yours to own.

How do AI showing tools prevent rental scams?

The most effective systems require identity verification — government-issued ID or bank-level verification — before issuing any lockbox code or showing access. A December 2025 FTC Data Spotlight documented nearly 65,000 rental-scam reports and approximately $65 million in losses since 2020; scammers specifically target the code-issuance step. An ID check after the fact, or an optional one, doesn't close that gap.

Does AI leasing software replace humans entirely?

No — and any vendor who implies otherwise is selling you a problem. The best systems handle high-volume, low-complexity interactions (inquiry response, FAQ, showing scheduling) autonomously, and have explicit rules for routing complex, edge-case, or high-intent conversations to a human, fast, with full context. Ask what specifically triggers that escalation.

What does "integrates with my PMS" actually mean?

It can mean anything from a one-way listing feed to a full two-way sync that writes applicant status and showing outcomes back into your system of record. A logo on a slide doesn't tell you which one you're getting. Ask for a live demonstration of what moves in both directions, and ask specifically whether the integration writes data back to your PMS or only reads from it.

How long am I locked into a leasing software contract?

Terms vary significantly, and the market's recent private-equity consolidation has made this question more important. Ask for the opt-out window in plain language, confirm that your pricing terms are in writing and survive a change of ownership, and verify that your data — contacts, showing history, applicant records — is portable when you leave.

What proof should a vendor show for my portfolio size?

Ask for named, specific results from a residential portfolio comparable to yours in door count and property type. Multifamily case studies don't translate directly to scattered-site or single-family residential operations. If a vendor can't point to a reference you can actually speak with, treat that as a signal worth taking seriously.

How fast should leasing software respond to a lead?

The MIT/InsideSales Lead Response Management study found that contacting a web lead within five minutes makes qualification roughly 21 times more likely than a 30-minute response, with lead quality dropping approximately 80% after that window. The best AI leasing tools respond in under a minute; ask each vendor what their documented response time is and how it's measured.

What if my leasing software vendor gets acquired?

Clarify this before you sign, not after. Ask whether your pricing is locked if the company changes hands, what your exit rights are, and who owns your data at that point. Operators who've been through PE-backed roll-ups in this market describe price increases and support degradation as the consistent pattern. A vendor with nothing to hide will put the protections in writing.

The best leasing software isn't the one with the longest feature list — it's the one whose vendor answers these questions without flinching. The demo shows you the happy path. These questions show you everything else.

Want to see what good answers to all seven questions look like? Book a LetHub demo.