FAQ
Questions worth answering before you talk to us.
Plain answers about what is measured, what is interpreted, and what nobody outside the model can promise.
- What exactly did you measure?
- 64 real buyer and seller questions for a defined market, each asked four separate times against ChatGPT (OpenAI) — up to 256 answers, captured verbatim and preserved. Each answer is classified for who was recommended as the person to hire, separately from who was merely mentioned. The valid-answer denominator is published with every benchmark.
- Can you guarantee that ChatGPT will recommend us?
- No. AI outputs are probabilistic, models change frequently, and we do not control their recommendation systems. We improve measurable inputs and representation, then track whether observable outcomes change.
- How is this different from SEO?
- SEO generally focuses on improving the discoverability and ranking of webpages. AI visibility examines whether AI systems understand and surface the brand itself across recommendation-oriented questions. The disciplines overlap, but the measurement problem is different.
- What would you actually change?
- It depends on the diagnosis. Typical work covers entity clarity (name, brokerage and profile consistency), third-party source coverage on the surfaces the answers actually cite, verifiable representation of production and record sales, and technical accessibility. Every intervention is logged with the evidence behind it, then remeasured with the same question set.
- Most of our business is referrals. Why does this matter?
- Referrals still close. But the referred seller now checks the referral before calling — and increasingly, they check with AI. When the answer confirms what their friend said, the call happens faster. When it names someone else, a warm introduction cools. AI doesn't replace your referral pipeline; it sits inside it.
- How do you know what influences an AI answer?
- We distinguish between what we can directly observe and what we infer. We capture model outputs, citations, sources, competitive patterns, entity information and other measurable signals. When causation cannot be established, we label conclusions as hypotheses rather than facts.
- What platforms do you measure?
- Our framework is designed to evaluate leading AI answer and recommendation platforms, with platform coverage determined by relevance, technical feasibility and the client's market.
- How long does this take?
- Measurement can begin immediately after onboarding. Changes in underlying signals and recommendation outcomes can occur on different timelines, so we track leading indicators as well as recommendation results rather than promising a fixed ranking deadline.
- Is this just content creation?
- No. Content may be one intervention, but the broader system can involve entity clarity, third-party authority, citations, structured information, technical accessibility, topical associations and other diagnosed gaps.
- Do we need to replace our SEO agency?
- Usually not. Strong SEO, PR, content and digital authority can support AI visibility. Our work is designed to identify the additional recommendation-specific gaps and coordinate with existing partners where useful.
- What access do you need?
- Access depends on the engagement. We minimize required permissions and define exactly what is needed during onboarding.
- What happens if our AI visibility is already strong?
- Then the audit should show that. We would rather identify that there is no meaningful problem than manufacture one.
- Do you know how ChatGPT or Google ranks sources?
- We do not reverse-engineer a private algorithm. We reverse-engineer the observable evidence environment around AI recommendations. We report what was asked, what was answered, what was cited and where evidence is missing.
- Do you publish client data?
- No, not without written permission. Benchmarks are published with entities anonymized, and every published figure traces back to a preserved answer we can show on request.
- Which AI systems do you measure?
- Current benchmarks are captured from ChatGPT (OpenAI). Coverage is stated on every report rather than implied, and additional providers are added only when access is stable enough to measure consistently.
Keep reading
See who AI recommends in your market — and where you stand.
One client per market. Start by telling me where you work.