Use case · Screened on every output
Put a Fair Housing screen between AI and publish
RealtrAI screens every client-facing output against federal protected classes and HUD advertising guidelines before the draft reaches you, so the listing copy, social post, or neighborhood guide you publish has already cleared a compliance review.
The challenge
- !One careless phrase about a school district, a church, or a quiet family street can read as steering, and you may not catch it.
- !Fair Housing language is easy to get wrong and the consequences land on the agent and the brokerage, not the tool.
- !Generic AI writers have no compliance layer, so they will happily produce copy that violates HUD advertising guidelines.
- !Manual review is inconsistent. What one agent flags, another ships, and there is no record of who decided what.
- !When a complaint comes in, you have no audit trail showing the language was reviewed before it went live.
How RealtrAI helps
- ✓RealtrAI screens every client-facing output against the 7 federal protected classes, Race, Color, National origin, Religion, Sex including sexual orientation and gender identity, Familial status, and Disability, plus state and local additions.
- ✓A pre-generation filter blocks prohibited language before the draft is written, so the model is steered away from problem phrasing from the start.
- ✓Every draft is reviewed against HUD advertising guidelines before it reaches your editor, with risky language flagged in context.
- ✓Each flag and each override is written to an audit log, so the brokerage has a record of what was caught and what was published.
- ✓Nothing auto-publishes. A human reviews the screened draft and makes the call, keeping a person in the loop on every piece of copy.
- ✓Screening covers 180+ jurisdictions, so the review reflects the protected classes that apply where the property actually sits.
How the screen works
A three-tier Fair Housing screen on every draft
This is not a disclaimer in the footer. The screen runs inside the generation path, so compliance is built into the output instead of bolted on after.
Pre-generation filter
Prohibited language is filtered before the draft is written, steering the model away from steering, exclusionary, and protected-class references from the start.
HUD-guideline review
Every output is reviewed against HUD advertising guidelines before it reaches your editor, with risky phrasing flagged so you can see what tripped the check.
Audit log
Every flag and every override is logged, giving the brokerage a defensible record of what was screened, what was changed, and what was approved.
Human in the loop
No client-facing copy auto-publishes. A person reviews the screened draft and decides, so judgment stays with the agent and the broker.
Jurisdiction-aware coverage
The screen accounts for state and local protected classes across 180+ jurisdictions, not just the seven federal classes.
Where it applies
The same screen on the copy that gets you in trouble
Listing descriptions, where a phrase about who a home is perfect for can read as a preference.
Social media posts, where short, casual copy is the easiest place to slip.
Neighborhood guides, where describing an area can drift into describing the people who live there.
Every client-facing draft, so the standard is consistent across the team instead of one agent at a time.
How it works
Adopting stay fair-housing-safe.
Generate from any client-facing tool
Write a listing, a social post, or a neighborhood guide. The Fair Housing screen runs automatically inside the generation, with no separate step to remember.
Prohibited language is filtered first
Before the draft is written, the pre-generation filter steers the model away from protected-class references, steering language, and exclusionary phrasing.
The draft is reviewed against HUD guidelines
The output is checked against HUD advertising guidelines and any state or local additions for the property's jurisdiction, and risky language is flagged in context.
You review the screened draft
The flagged draft reaches your editor. You read the flags, adjust the copy, and a human makes the final call before anything publishes.
Every flag and override is logged
The audit log records what was caught, what you changed, and what you approved, so the brokerage has a clear trail if a question ever comes up.
FAQ
Questions, answered.
Does RealtrAI screen every output or just listings?
Every client-facing output is screened, including listing descriptions, social posts, and neighborhood guides. The Fair Housing screen runs inside the generation path, so it is not something an agent can skip.
Which protected classes does the screen cover?
The seven federal protected classes: Race, Color, National origin, Religion, Sex including sexual orientation and gender identity, Familial status, and Disability, plus state and local additions across 180+ jurisdictions.
How does the three-tier screen work?
First, a pre-generation filter blocks prohibited language before the draft is written. Second, the output is reviewed against HUD advertising guidelines before it reaches your editor. Third, every flag and override is written to an audit log.
Does RealtrAI publish copy automatically?
No. Nothing client-facing auto-publishes. A human reviews the screened draft and decides what goes live, so a person stays in the loop on every piece of copy.
Does this replace legal review or my broker's compliance process?
No. The screen reduces risk and gives you an audit trail, but it does not replace your brokerage's compliance process or legal counsel. Fair Housing screening in RealtrAI is governed by Trunnion AI.
The tools that get you there
Get started
Put this outcome to work.
Try the tools free for seven days. No credit card.