1. Establish the baseline
Agree on one location, one priority service, three search surfaces, and ten buyer questions before work begins. A surface means a particular product and mode, such as Google Search, ChatGPT with search, or Claude with search. Record language, location context, date, and exact wording. Availability and supported data sources are agreed during onboarding.
2. Preserve observations
For each question, record the answer, sources, named businesses, and any uncertainty. Use two runs per question per surface where access permits. If a check fails or a surface is unavailable, label it unavailable. Do not count an unavailable check as an absence. Answers can vary across accounts, location, product versions, and time.
3. Report the denominator
If we report mention frequency, the denominator is the number of successfully completed observations in the agreed sample. It is not the share of all customers, all AI searches, or the market. For example, being named in 6 of 20 valid observations means 30% of that sample only. This is an explanation, not an FLP client result.
4. Keep a record of work
Every delivered change includes the source of the problem, proposed correction, necessary approval, completion date, and evidence such as a page URL or before-and-after record. Profile changes may remain pending external platform approval. Preparing a change and having it accepted are different statuses.
5. Measure business outcomes separately
With authorized account access, establish website inquiry events and use available call or booking reports. Agree on what makes an inquiry qualified. Report unique inquiries, appointments, and won work only when the underlying records support them. A website click or AI mention is not a booked job.
6. Explain uncertainty
Seasonality, competitors, advertising, pricing, and response speed can affect outcomes. A before-and-after change is not by itself proof of causation. Google Search Console includes its AI Search features within overall Web performance; that does not provide a clean platform-wide AI recommendation count. Google’s measurement guidance.
What the free brief builder measures
Nothing about your live search position. It produces a rule-based starting plan from your answers. It does not crawl your website, connect accounts, submit changes, or claim a visibility score. The interactive homepage workspace is also an explicitly labeled example.