Skip to content

Last updated October 21, 2026

AI Disclaimer

What Real Extent's AI can and can't do, and why you stay in charge of what it produces. This disclaimer is part of our Terms of Service.

1. AI output can be wrong

Real Extent uses artificial intelligence to read, summarize, sort, draft, and suggest. Its output, including email drafts, summaries, lead scores, meeting briefs, contract dates, listing descriptions, price estimate talking points, and answers about your documents, can be inaccurate, incomplete, out of date, or unsuitable for your situation.

Review anything the assistant produces before you rely on it or share it.

3. You're responsible for what goes out from your accounts

You are responsible for every communication sent from your email or calendar through Real Extent, including messages and invitations sent automatically by actions set to Auto. Calendar events created by the assistant may send invitations to their guests.

Review your permission settings regularly, and use Ask first for anything you want to check before it goes out.

Where the law requires you to tell people that AI was involved in a communication, you are responsible for doing so.

4. Safeguards don't replace your review

By design, the assistant never sends bulk email, cancels meetings, or sends contracts or legal documents on its own; doesn't agree to prices, commissions, or other financial terms; and treats instructions found inside emails and documents as untrusted. These safeguards reduce risk but can't remove it.

5. Lead scores, dates, and price estimates

Lead scores are estimates based on signals such as timeline, pre-approval, and engagement. They don't predict whether someone will buy or sell, and must never be used to treat people differently because of a protected characteristic.

Dates found in a contract can be missed or misread. Always check them against the signed contract. You are responsible for every deadline.

Price estimates are calculated from the comparable sales you provide. They are not appraisals, and their quality depends on the data you enter.

Document summaries can miss important details. Read the original document, such as the full inspection report, before advising a client.

6. AI models we use

The assistant uses Claude models from Anthropic. Search over documents and saved memories uses OpenAI embeddings. Model behavior can change over time, and the same request can produce different results. When your AI allowance runs out, the assistant uses simpler built-in rules, which are more limited.

7. Reporting problems

If the assistant produces something harmful, inaccurate, or unexpected, tell us at contact@aura-technologies.co. You can review every action it took in your activity log.

Questions? Contact contact@aura-technologies.co.