Home › Methodology

Methodology & pledge

PropCite Citation Index v1

This page explains exactly how PropCite measures AI citation share, and the commitments that keep the index neutral. Trust is the product, so the method is public and dated.

How citation share is measured

Home sellers increasingly start with an AI assistant — "Who's the best listing agent in my city?" — instead of a web search. Whichever agents those assistants name win the introduction. PropCite measures exactly that, and nothing else.

  1. Engines. Each cycle we query three AI assistants: ChatGPT, Claude, and Google AI.
  2. Questions. For every city we cover, we ask the same standardized pack of 50 seller-intent questions — the real questions a home seller would ask before choosing who lists their home.
  3. Tally. We read each AI answer and record every agent named in it. Agents named in the answers are tallied across all 50 questions and the engines above.
  4. AI answer share. An agent's score is the percentage of those answers that name them. Their rank in a city is their share relative to other named agents.
  5. Threshold. We publish agents named in at least two answers, so leaderboards reflect repeated, durable citations rather than a single one-off mention.
  6. Cadence. Rankings are updated monthly. Each cycle re-asks the full 50-question pack per city across the three engines.

That's the whole signal: AI citation share across the engines sellers actually use, refreshed monthly. As a worked example, the agent named in the most answers in a city is shown as No. 1 by AI answer share.

No-paid-placement pledge

Rankings are earned, never bought. An agent cannot pay to move up, and money never moves a ranking. There is no paid placement anywhere on PropCite.

Claiming a profile is free and does not change where an agent appears. We add claimed details (bio, brokerage, contact) only with the agent's consent, separately from the AI-measured ranking — they never affect the score.

Disclosed ownership

PropCite is operated by the team behind DoIAppear, an AI-visibility diagnostic for real estate professionals. We disclose this on purpose: transparency is part of what keeps the index trustworthy.

DoIAppear cannot influence PropCite rankings. The index is computed purely from what AI assistants say across the 50 questions. No DoIAppear customer can buy a position, and paying for help does not move a ranking — it can only make an agent genuinely more citable through real content, reviews, and structured data, which the engines may or may not then reflect.

Rankings are earned, not bought — PropCite earns nothing from rankings and DoIAppear cannot influence them.

Changelog

PropCite Citation Index v1 — initial public release
Three engines (ChatGPT, Claude, Google AI), 50 seller-intent questions per city, monthly cadence, two-mention publication threshold. This is the first published version of the index; future revisions and their dates will be listed here.