Last updated: September 12, 2026
Fair housing applies to the phone
Fair housing rules apply to a leasing phone call the same way they apply to a tour. Risk shows up when an agent asks one caller about children and another caller nothing, or answers an availability question differently depending on the voice on the line. Consistency across callers is the part an operator can measure and evidence afterward.
Every property management company in the country trains on this, most of them once a year, and the better programs use real scenarios instead of a slide deck. None of it reaches a leasing agent at 4pm on a Tuesday, forty seconds into an inquiry call, when a caller mentions her twins start at the school down the road and asks whether the building stays quiet.
What the agent says in the next ten seconds is where the exposure sits. "We have a lot of families here" is one answer. "Most of our families are on the first floor, it gets quieter upstairs" is another, said just as warmly by the same person, and it has moved a caller toward part of a building for a reason the rules name.
The phone is covered because HUD wrote it that way. At 24 CFR 100.75(b), the prohibitions on discriminatory notices and statements reach written or oral statements by a person engaged in the sale or rental of a dwelling. A sentence said out loud to a caller is a statement.
Where the risk lives in a leasing call
The Fair Housing Act names race, color, religion, sex, familial status, and national origin in 42 U.S.C. 3604(a) and (b), and adds handicap through subsections (c) into (f). States and cities add categories on top of the federal list, so what your agents work under is longer than what the Act says, and your counsel is the one who tells you how much longer.
Going statute by statute produces a decent outline and a useless training tool, so the order below is the order things happen on a call. Steering generates most of the phone exposure, followed by questions that touch familial status, with disability inquiries close behind and easier to walk into than either.
The greeting and who the caller sounds like
A caller with an accent gets a shorter call. The name an agent struggles with gets less patience on the spelling. Nobody would know, because the agent who does it least sits four feet from the agent who does it most and neither has ever heard the other work.
The regulation at 24 CFR 100.70(a) reaches conduct that restricts or attempts to restrict a person's choices by word or conduct. A call that gets less effort is a thin version of that, and whether a particular call crosses a line is a question for a lawyer rather than a scorecard.
Attentiveness also gets talked about more than it gets measured, which is worth being careful about. A criterion flagging an agent for sounding tired at 4:40 gets ignored by week three, so the useful version is a pattern across fifty calls.
Qualifying questions asked unevenly
Familial status is where the phone gets operators into trouble, mostly through friendliness. An agent hears kids in the background, asks how old they are, and that call now carries a question the last four calls did not.
Disability is the tighter one. Under 24 CFR 100.202(c), it is unlawful to inquire whether an applicant has a handicap or to inquire into the nature or severity of one, with four narrow exceptions, each holding only where the question goes to every applicant. The trap is sympathetic rather than hostile: a caller says the stairs are hard on her, and the agent asks whether that is something the property needs to accommodate, while trying to help.
Uneven is doing a lot of work in that heading, and it is worth slowing down on. A caller who volunteers that her twins start school in the fall, followed by an agent who says that's great and moves on, is a different record than an agent asking a cold caller how many children will live in the unit. Both produce a mention of children in the transcript. One of them is a leasing agent being a person.
A fixed question set the agent works from on every call is the fix for the second one, and the qualifying questions in a leasing script covers what safely goes to every caller.
Availability and pricing answered differently
Availability carries the clearest statutory hook. Under 42 U.S.C. 3604(d), representing that a dwelling is not available for inspection, sale, or rental when it is in fact available, on a protected basis, is prohibited outright.
The softer version is more common and harder to catch. HUD's rule at 24 CFR 100.70(c) reaches discouraging somebody by exaggerating drawbacks or by failing to mention desirable features, which in a leasing office looks like the two bedroom on the third floor coming up for one caller and never for the next. The same regulation reaches assigning people to particular sections or floors of a building, which is the polite sentence from the top of this page.
What happens after the call
Consistency does not stop when the agent hangs up. One caller gets a callback in ninety minutes, another gets an email three days later, and neither agent could tell you why. Delay in processing an application sits in 24 CFR 100.70(d) alongside the rest of it, and your CRM already holds the timestamps nobody has ever laid beside the calls that produced them.
What training covers and what it leaves open
Annual training does a job worth paying for. It teaches the categories, gives agents scenarios with the right answer attached, tells a new hire who to escalate a reasonable accommodation request to, and produces a signed record.
The better training vendors have written about listening bias specifically, the Fair Housing Institute among them, so the industry already knows the phone is a soft spot. Observing it is a different budget from knowing it.
Nothing inside a training program listens to a call. The program ends, everyone goes back to the desk, and the next eleven months run unobserved until a complaint arrives. That gap is not a training failure, since training was never the instrument for this.
Proving consistency after the fact
Consistency is measurable in a way intent never will be. Say one property takes 60 inquiry calls a month across three leasing agents. Score all 180 against one rubric, hold the qualifying question set as its own criterion, and the open question stops being whether anybody treats callers differently and becomes which callers got asked what.
Mystery shopping is the incumbent answer, covering a handful of staged calls per property per year against the real inquiry volume, which is why scoring every inquiry instead of a sample produces a different record. A shop tests the script. Scoring a month of calls tests the floor.
The leasing call scorecard carries the consistency criteria as weighted lines. A criterion written as a fixed phrase, or as a question that was either asked or not, scores far more reliably than one asking whether an agent sounded engaged, which Aircall and StackAI both note about automated QA generally.
One limit bites harder here than anywhere else on this site. Speaker attribution on mono or non channelized audio can put a line on the wrong person, and every criterion above turns on whether the agent said something or the caller did. A caller volunteering that she has two kids and an agent asking about them are opposite records, and a transcript that swaps the two produces a confident wrong score on the exact criterion the program was built for. Check what your phone system exports before building a rubric on top of it, and while that defect is open, treat speaker dependent criteria as reviewer work with the transcript as a finding aid.
Scoring also says nothing about why two calls went differently. A gap between two callers is something to look at, and a lawyer decides what it means. What a supervisor gets from CallSniper is flagging a critical fail the same day, with a transcript timestamp under each flag, instead of finding out in a deposition. Call QA for property management holds the rest of that program.
Recording and disclosure obligations that come with this
A page telling you to score every leasing call owes you the obvious complication. Recording inbound calls brings its own consent rules, and several states require every party on the line to agree rather than just one.
Your disclosure wording and the states your properties operate in get decided before the first recording exists, and call recording consent by state covers that ground. Settle it first, because a compliance program built on recordings you were not entitled to make is worse than the gap it was meant to close.
What this page is not
None of this is legal advice, and nobody at Trinsic is your lawyer. Everything above describes rule text anyone can read, plus a way to observe your own phones. Your counsel decides what applies to your properties and what any particular call means.
The two records that hold up
Consistent handling is the defense. Paperwork showing you handled callers consistently is what turns that from a belief into a record.
A training completion record proves one thing: your company told its agents the rules on a date, and here are the signatures. That is intent, and intent counts for something. A scored recording proves practice, because it shows what an agent said on a Tuesday in March rather than what the agent was taught in January.
The second one changes the conversation with counsel because it arrives before the complaint does. A lawyer working from a training binder is building an argument. Four hundred scored inquiry calls, question sets side by side and the recording attached to each score, is a record that was produced on a schedule with no incident in view. Keep the scores whether they look good or not, since a file holding only the clean months invites its own question.
Start with one property and twenty recent inquiry calls. Lay the qualifying question sets next to each other, agent by agent, and see whether the same four questions got asked twenty times. The paid pilot runs six months of your own leasing audio against the rubric you approve and ends in a 30 minute working session on those calls, at a fixed pilot price.
Frequently asked questions
What questions are illegal to ask a potential tenant?
The Fair Housing Act at 42 U.S.C. 3604 prohibits discrimination on race, color, religion, sex, familial status, national origin, and handicap, and 24 CFR 100.202(c) separately prohibits inquiring whether an applicant has a handicap or into its nature or severity, with four narrow exceptions. State and local law adds categories on top. Which questions your agents can ask is a decision for your own counsel.
Does fair housing apply to phone calls?
Yes. HUD's rule at 24 CFR 100.75(b) applies the prohibitions on discriminatory notices and statements to written or oral statements by a person engaged in the sale or rental of a dwelling. The difference from a tour is that the call leaves a recording.
How often is fair housing training required for property managers?
Requirements vary by state, by employer policy, by any designation your team holds, and by the terms of a settlement or funding agreement. No single national rule sets a frequency for everyone, and annual is common practice rather than a requirement.
Can you record leasing calls?
Consent rules vary by state, and several require all parties on the line to consent rather than one. Settle your disclosure wording and your state coverage before recording anything, and call recording consent by state goes through it.
How do property managers prove fair housing compliance on calls?
Consistent handling across callers is the part that can be evidenced. Score every inquiry call against one rubric, hold the qualifying question set as a criterion so it compares caller to caller, and keep the recording attached to each score. Training records show what agents were told, and scored recordings show what agents did.
Sources
- 42 U.S.C. 3604, protected characteristics and the availability provision at 3604(d). Primary text, read 2026-09-12 via the Cornell mirror. Verify against uscode.house.gov.
- 24 CFR 100.70, 100.75, and 100.202(c). Primary text, read 2026-09-12 via the Cornell mirror. Verify against ecfr.gov.
- 42 U.S.C. 3602(k), familial status definition. Not read for this draft, confirm before publish.
- State and local additions to protected class lists. Unverified, varies by jurisdiction.
- Fair Housing Institute material on listening bias. Third party, unverified.
- Mystery shopping as the incumbent method for multifamily call quality. Client research, 2026-09-08, unverified.
- Two party consent states. Unverified here, confirm against the recording law page.
- Aircall and StackAI on rule based criteria scoring more reliably than tone. Third party, unverified.
- Transcript timestamps, confidence scores, critical fail flags, manual upload. CallSniper product documentation. Verified internally.
- Speaker attribution on single channel recordings. Diarization is less reliable when both sides of a call share one channel, which is why the scoring section tells you to check what your phone system exports.
- Oral statements and the reach of the Fair Housing Act, 24 CFR 100.75(b). Read via the Cornell LII mirror on 2026-09-12 because ecfr.gov refused the request. Confirm on eCFR before publish.
- Disability related inquiries and their exceptions, 24 CFR 100.202(c). Read via the Cornell LII mirror on 2026-09-12. Check against eCFR before publish.