AI Transparency Notice

This website

The public RealtrAI marketing website does not provide an AI chatbot, generate personalized housing content, score visitors, rank applicants, or make consequential decisions. Public calculators perform user-controlled arithmetic. Identifier-free aggregate page-load counts run until rejected; detailed analytics requires acceptance.

RealtrAI product workflows

RealtrAI may assist with drafting marketing copy, client communications, presentations, professional profiles, research questions, scenario practice, and calculations based on user-supplied assumptions. RealtrAI is not intended to profile a housing consumer or make or substantially facilitate an automated decision with a legal or similarly significant effect on a housing consumer. It must not rank applicants, determine eligibility, steer a person toward or away from housing, or replace a licensed professional's judgment. Exact workflows, provider, permitted inputs, retention, training-use controls, output handling, and availability must be confirmed in the current product release and written agreement.

Fair Housing and human review

Relevant workflows include Fair Housing-aware instructions, but the current product does not claim an independently verified automated Fair Housing screen, protected-class decision engine, targeting or ranking safeguard, local-law determination, or compliance audit log. Outputs can be inaccurate, biased, incomplete, outdated, or unsuitable. A licensed professional must review every draft against property facts, brokerage policy, MLS rules, federal Fair Housing requirements, additional state and local protected classes, and qualified advice before publication or client use.

Model training and customer content

The three questions every vendor review asks, answered plainly:

• Is customer content used to train models for other customers? No. Customer inputs, prompts, uploaded content, and generated outputs are never used to train or improve models across tenants. One customer's content does not shape any other customer's results.
• Is customer content used within your own workspace? Customer content may be used to improve results inside that customer's own workspace only - tenant-scoped adaptation such as saved context and preferences that shape that workspace's drafts. It is governed by the written agreement, and it never crosses into another tenant or a shared model.
• Can a third-party model provider train on customer content? RealtrAI uses contracted foundation-model provider(s) to serve generation requests; the provider(s) may vary by product release. The contracted provider(s), their retention and training controls, and the commitments that apply to your data are identified in the written agreement and data processing terms - confirm them there before production use. The current provider dependency list is on the subprocessors page.

Can tenant-scoped adaptation be disabled? A generally available self-service disablement control is not currently claimed. A customer may request an adaptation restriction or disablement by emailing contact@trunnion.ai. The response must identify the affected workspace, effective time, technical scope, residual saved context, and verification evidence. Do not submit production content until the restriction is confirmed in writing and accepted for the deployed release.

Output provenance

RealtrAI does not currently claim that exported drafts carry a persistent watermark or machine-readable AI marker. A licensed professional who approves and publishes a draft takes responsibility for the final communication. Material AI involvement should be disclosed where a client, consumer, regulator, or counterparty could otherwise reasonably infer independent human authorship or professional review.

Data handling boundaries

Do not enter sensitive, confidential, protected-class, client, listing, financial, or personal information unless the written scope and deployed controls expressly permit it. Where processing occurs and the exact provider configuration for your deployment must be confirmed in the written agreement before production use.

Monitoring, complaints, and escalation

Customers remain responsible for assigning human owners, validating configuration, monitoring outcomes and drift, reviewing complaints and corrections, and completing any required impact or Fair Housing assessment. Product use must include a path to stop, correct, or escalate concerning output. A consumer or professional may ask for human review or correction of an output attributed to a RealtrAI draft by emailing contact@trunnion.ai. Include no sensitive client or housing information in the initial email.