> For the complete documentation index, see [llms.txt](https://docs.layerlens.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.layerlens.ai/more-in-this-section-2/scenarios-9.md). # Real estate and PropTech scenarios Five canonical scenarios where brokerages, MLS operators, lenders, landlords, and PropTech platforms use Stratix evaluation to gate AI features. ## 1. Listing assistance and content generation **Setup:** AI generates listing copy, summarizes neighborhood features, and surfaces comparable properties. Agent reviews; AI doesn't publish unsupervised. **Why it's hard:** * Fair-Housing-Act-protected categories cannot be referenced even indirectly (no "perfect for young families," no "walking distance to St. Mary's") * Neighborhood characterizations are landmines — historical bias surfaces fast * Inaccurate descriptions are commonly litigated as misrepresentation **Stakes:** * Fair Housing Act violations carry civil + criminal liability * HUD has investigated AI-driven listing language specifically * State broker licensing boards can suspend / revoke licenses * DOJ pattern-or-practice suits against brokerages have followed AI listing tools **Stratix evaluation criteria:** | Aspect | What you measure | Pattern | | ------------------------ | ------------------------------------------------------------------ | ----------------- | | Protected-class language | No protected references direct or indirect | Hard rule + judge | | Steering language | No language steering toward/from neighborhoods on protected basis | Hard rule + judge | | Factual accuracy | Square footage, beds, baths, year-built match listing data | Code assertion | | Source-citation | Neighborhood claims cite source (school district data, walk score) | Citation rule | | Agent-review gate | No publish without agent sign-off | Hard rule | ## 2. Property valuation (AVM-style) **Setup:** AI generates property valuation estimates from market data, comparables, property attributes. Output: estimated value, confidence interval, comp set. **Why it's hard:** * Bias against neighborhoods of color is well-documented in appraisal industry; AVMs inherit data bias * Federal regulators (PAVE task force) have AVM bias as a priority * ECOA and Fair Housing Act apply to lender use of valuations **Stakes:** * Lender-reliance liability for biased valuations * Class-action exposure for systemic undervaluation in protected neighborhoods * CFPB and HUD have proposed AVM rules under FIRREA Title XI **Stratix evaluation criteria:** * **Demographic fairness scorer** — Valuation parity across neighborhoods after controlling for property attributes * **Comp-set transparency rule** — Every valuation cites the comp properties * **Confidence-interval rule** — Required band; missing CI fails the trace * **Out-of-area refusal rule** — Hard rule: AVM refuses where data quality is below threshold * **Audit trail per valuation** — Inputs, model version, output retained ## 3. Tenant screening **Setup:** AI scores rental applicants on credit, eviction history, criminal records, income. Property manager decides; AI does not unilaterally accept/deny. **Why it's hard:** * FCRA applies if it's a consumer report — strict accuracy and dispute rights * Disparate impact on protected classes — well-documented in criminal-record use * HUD 2016 guidance on criminal-record use in housing decisions * State and local "ban the box" / source-of-income laws **Stakes:** * FTC and CFPB enforcement on tenant-screening companies * Class-action exposure under FCRA + FHA * HUD complaints + state/local discrimination suits **Stratix evaluation criteria:** * **FCRA accuracy rule** — Hard rule: every reported item must be verifiable * **Adverse-action notice rule** — Output triggers FCRA adverse-action notice when used to deny * **Disparate-impact monitoring** — Group fairness on outcomes * **Criminal-record-rule compliance** — Hard rule per HUD guidance: arrest records excluded; convictions weighted by recency, severity, relevance * **Source-of-income rule** — Hard rule where applicable per state/local law ## 4. Transaction support and document review **Setup:** AI summarizes purchase contracts, surfaces inspection issues, drafts disclosures, flags missing documents. **Why it's hard:** * Real-estate contracts are state-specific; clauses interpret differently across jurisdictions * Disclosure obligations vary (CA Natural Hazard Disclosure, lead-paint federal, etc.) * Missed inspection issues become latent-defect claims **Stakes:** * Failure-to-disclose suits are a primary RE litigation category * E\&O insurance excludes some AI-driven errors * Closing-time issues compound into transactional collapse **Stratix evaluation criteria:** * **State-specific clause rule** — Hard rule: contract analysis matches the applicable state's standard form * **Required-disclosure rule** — Hard rule: federal lead-paint, state natural-hazard, etc. surfaced when applicable * **Inspection-issue completeness** — Code assertion: high-priority items surfaced * **Citation faithfulness judge** — Output grounded in cited contract section ## 5. Property management and tenant communications **Setup:** AI handles routine tenant communication: rent reminders, maintenance request triage, lease-question responses. **Why it's hard:** * Tenant rights vary by jurisdiction (rent control, just-cause eviction, security-deposit rules) * Habitability standards are non-negotiable * Retaliation and discrimination claims attach quickly **Stakes:** * State and local fair-housing law (often stronger than federal) * HUD investigation triggers from systemic complaint patterns * Rent-control violations carry per-tenant damages **Stratix evaluation criteria:** * **Tenant-rights rule** — Hard rule: outputs reference the applicable jurisdiction's tenant rights when relevant * **Habitability rule** — Maintenance issues affecting habitability auto-escalate * **Retaliation guard rule** — Outputs in tension with a recent tenant complaint route to human * **Tone judge** — De-escalation quality * **Multilingual parity** — Same accuracy across language pairs ## See also * [Real estate eval patterns](/more-in-this-section-2/eval-patterns-9.md) * [Real estate compliance](/more-in-this-section-2/compliance-9.md) --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://docs.layerlens.ai/more-in-this-section-2/scenarios-9.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.