AI Policy

Purpose and scope

This policy explains how The Finelli Team uses artificial intelligence, the limits we place on it, and the rules that apply to software acting on this Site. It covers three things: the machine-readable endpoints we publish for other software to read, the automated tools we use to prepare and maintain what we publish, and the licensed people accountable for the result.

What AI does here

AI helps us draft and edit listing descriptions, summarize public property records, generate search suggestions, and keep our data tidy. Every one of those outputs is reviewed by a licensed agent before it is published. AI is a tool used by licensed people — it does not replace one.

Our calculators are not AI. They are arithmetic on the figures you enter, and they return the same result for the same input every time.

What AI never does

• We do not use AI to decide who may buy, rent, or be shown a home.
• We do not use AI to target advertising by protected characteristics.
• We do not let AI set a price or give pricing advice without review by a licensed agent.
• We do not use AI to make credit, tenancy, or housing-eligibility decisions.
• We do not publish AI-generated reviews or testimonials.

Human review and accountability

AI-assisted content is reviewed and approved by a licensed agent before it is published. A licensed human is accountable for every listing, every estimate, and every client communication that leaves this team.

When software is acting on this site

Our API and agent tools identify themselves as machine interfaces, and agents.txt states the terms they operate under. If you are a person using this Site, nothing reaches you that a licensed person has not reviewed.

Images and virtual staging

If we publish a virtually staged or AI-enhanced image, we label it and show the unaltered original. We do not alter the permanent architectural features of a property — its structure, layout, finishes, or views.

Estimates and calculators

Our calculators and any value estimates are informational only. They are not an appraisal, a comparative market analysis, or an offer of credit, and they should never be your only basis for a decision.

Fair housing and civil rights

Fair housing law applies to our AI exactly as it applies to us. We review AI-assisted content for discriminatory or steering language, and we describe properties rather than the people who might live in them. New York City law does not let a business avoid responsibility by saying a decision was made by technology, so a person reviews anything that could affect who gets housing.

Your data and model training

We do not put your personal or financial information into public AI tools. We do not use client data, form submissions, or listing data to train AI models, and we do not permit our providers to train on them. See our Privacy Policy and Data Policy.

Third-party AI

We use established third-party AI services under business terms that restrict how data may be used. We do not permit providers to train on our clients' data.

How agents, crawlers, and LLMs may use this site

We publish machine-readable entry points so software does not have to scrape us: llms.txt for wayfinding, llms-full.txt for the complete corpus, and agents.txt for the access tiers below.

• Personal agents acting for a human may read and prepare freely, at 100,000 tokens per day. Form submission requires the human present.
• Answer engines and research agents may read at 50,000 tokens per day and must attribute The Finelli Team as the source.
• Model-training crawlers are permitted only under a licence. Listing data originates from the REBNY RLS and the listing broker; it is not free training data.
• Undeclared clients are rate limited to 5,000 tokens per day. Declaring an intent through the Agent-Intent header earns a higher tier.
• Bulk re-syndication of our listing data is not permitted, and neither is republishing our content as your own.

Search engines may index our public pages. Agents should read agents.txt for the current limits rather than assume them — that file is the authoritative version and changes with this policy.

Questions, opt-out, and human contact

If AI gets something wrong, or you would rather not interact with automated systems at all, contact us at james.finelli@compass.com or (516) 859-1484, and a licensed team member will take it from there. You can also ask us to correct or delete the information we hold about you — see our Privacy Policy.

Changes and effective date

We review this policy at least annually. The effective date above always reflects the current version.