AI visibility for real estate

When buyers ask AI who to hire, is your team in the answer?

Recommended First measures which real estate teams AI recommends, compares those recommendations with verified production, and identifies the evidence and citation gaps behind the answers.

Prompts
64 buyer and seller questions
Repetitions
4 per question
Provider
ChatGPT (OpenAI)
Capture date
August 2026
Valid answers
Published with each benchmark (256 maximum)

Real-world production and AI visibility are two different things.

A team can rank near the top of its market by closed volume and still be absent from the answers buyers read first. Production is recorded in MLS systems. AI answers are assembled from the public information environment — portal profiles, brokerage pages, local publications, review platforms, forums. When those two records disagree, the answer follows the public record, not the production record.

We do not reverse-engineer a private algorithm. We reverse-engineer the observable evidence environment around AI recommendations.

Anonymized measured example · Jersey City, NJ · residential
ProductionMarket rankAnswers recommending
The team (you)$25.4M closed#8 by volume4 / 256
Competitor$17.1M closedUnranked116 / 256

Illustrative layout. Public benchmarks use verified production data and captured answers.

Mentioned is not the same as recommended.

An answer that lists your brokerage, or names you in passing while recommending someone else, is not a recommendation. We classify every captured answer separately: recommended as who to hire, mentioned without a recommendation, or absent. Only the first counts toward recommendation frequency.

One answer is also not a measurement. AI outputs vary between runs, so a single screenshot proves nothing about the pattern. Each question is asked repeatedly, every response is preserved word for word, and every figure is reported with its denominator.

What we measure

Recommendation frequency
How often the team is named as who to hire, across every captured answer.
Competitor presence
Which teams are recommended instead, and how often each one appears.
Question-level performance
Which buyer, seller and neighborhood questions the team appears in, and which it never does.
Citation and source coverage
Which domains are cited around the answers, and whether the team appears on them.
Entity accuracy
Whether the name, brokerage, markets and team members are consistent across public sources.
Description accuracy
Whether the team is described correctly when it does appear.
Verified production comparison
Measured recommendation frequency set against verified closed production.

Measure, compare, diagnose, improve, remeasure.

  1. 01

    Measure

    Run a fixed question set repeatedly and preserve every answer verbatim.

  2. 02

    Compare

    Set recommendation frequency against verified production and against competitors.

  3. 03

    Diagnose

    Identify the evidence and citation gaps around the answers.

  4. 04

    Improve

    Make legitimate corrections to the public evidence environment.

  5. 05

    Remeasure

    Rerun the same locked question set under comparable conditions.

What this cannot prove

  • That any specific change caused a change in AI answers.
  • That the answers captured will be the answers produced tomorrow.
  • How a model ranks or weights any source internally.

See who AI recommends in your market — and where you stand.

One client per market. Start by telling me where you work.