How FirstPick measures your AI Visibility Score

Your AI Visibility Score is the percentage of sampled AI answers about your services in which your business is named. It is a measurement of what AI assistants actually say, not a prediction and not a promise.

A second, separate number appears beside it throughout your report: share of named mentions is your share of every time AI named any business in the answers we sampled, split between you and the competitors named alongside you. Different denominators, different questions. We label which one you are reading, every time.

What we measure

For each business, FirstPick builds a basket of buyer-intent questions: the questions real customers ask when they are choosing a provider, phrased across your service lines and neighborhoods (for example, “best med spa near South Congress” or “who should I see for veneers in Denver”). We put those questions to the assistants we measure, geo-anchored to your market, and record which businesses each answer names.

• Coverage today and next: We check Google AI Overviews, Google AI Mode, Gemini and ChatGPT. Your AI Visibility Score is computed from Google AI Overviews, Google AI Mode, Gemini and ChatGPT. This same list appears everywhere on the site, so what we market always matches what we measure.
• When an assistant is quiet: We run each scan across the AI assistants and report what they say, so results vary and we can't promise a specific outcome. When a scan has enough data we send your report, building the score from the assistants that answered and noting any that didn't. When a scan doesn't meet our quality bar, we review it before sending rather than hand you a thin result.
• Repetition: AI answers vary, so each question runs multiple times and we score across the repetitions, never off a single answer.
• Slices: results are scored overall, per assistant, and per service line, so you can see exactly where the gap is. Where we blend the assistants together we say so, and we show the per-assistant split beside the blend.
• Sampling: recurring monitoring samples a rotating subset of the basket each cycle. It is a sample, not a census, and we label it that way.

What we inspect before recommending fixes

To recommend fixes, FirstPick reviews public information only: the business website, visible service pages, visible structured data, public business-profile information, and sources that appeared in captured AI answers. We do not access analytics, booking systems, ad accounts, patient data, or customer records. Recommendations are based on observed gaps and public source patterns, not a guarantee of why an assistant chose one business over another.

What we do NOT claim or measure

• We never access your website analytics, booking system, or any private customer data. We observe public AI answers and read public pages.
• We never guarantee rankings or results. Nobody controls what an assistant says; we measure it and recommend fixes that improve the source material assistants read.
• Our revenue figures are opportunity ranges built from stated assumptions you can see and adjust. They model opportunity exposure; they do not promise recovered revenue.
• Movement in your numbers is directional evidence, not proof of cause. We report what changed and when; we do not claim credit for any single change.

Common questions

What is my AI Visibility Score?

Your AI Visibility Score is the percentage of sampled AI answers about your services in which your business is named. If we sample 100 AI answers about your services and your business is named in 12 of them, your AI Visibility Score is 12%.

How is that different from share of named mentions?

Share of named mentions is your share of every time AI named any business in the answers we sampled, split between you and the competitors named alongside you. The two answer different questions and move independently. Your AI Visibility Score counts answers: of the answers we sampled, how many named you. Your share of named mentions counts names: of every business AI named across those answers, how many of those names were you. If AI names five businesses in every answer and you are always one of them, your AI Visibility Score is 100% and your share of named mentions is 20%.

Why do you repeat the same question multiple times?

AI answers vary from run to run. A single answer proves little, so FirstPick asks each question several times and scores your AI Visibility Score across all the repetitions rather than trusting any single result.

What does "named on at least one surface" mean?

We measure several AI surfaces, and they disagree with each other constantly. "Named on at least one surface" is the percentage of ALL the questions we tracked, including the ones that name your business, where at least one measured AI surface named you, counted once per question. It counts a question as won if any one surface named you, so it reads higher than any single surface on its own. That is why we show it beside the per-surface split rather than instead of it: the blend tells you whether buyers can find you at all, and the split tells you where.

Does FirstPick access my website, analytics, or customer data?

No. FirstPick only observes what public AI assistants say and reads public pages. We never connect to your analytics, booking system, or customer records, and we never install anything on your systems.

Can FirstPick guarantee that AI will recommend my business?

No, and we say so everywhere. Nobody controls what an AI assistant says. FirstPick measures where you stand, identifies the gap, and generates fixes that give assistants clearer source material. Results are directional, never guaranteed.

Which AI assistants does FirstPick measure?

We check Google AI Overviews, Google AI Mode, Gemini and ChatGPT. Your AI Visibility Score is computed from Google AI Overviews, Google AI Mode, Gemini and ChatGPT.

What if an assistant does not return enough answers about my business?

We run each scan across the AI assistants and report what they say, so results vary and we can't promise a specific outcome. When a scan has enough data we send your report, building the score from the assistants that answered and noting any that didn't. When a scan doesn't meet our quality bar, we review it before sending rather than hand you a thin result.

Why do my results change month to month?

AI answers can vary over time, and between two repeated questions, so some movement is normal. That is why we repeat the measurement rather than checking once: a single snapshot is directional, and the history is what shows whether a pattern is holding or genuinely changing.

For the broader questions owners ask before this level of detail, the learn guides cover what AI visibility is and how assistants choose the businesses they name.