.svg?fit=max&auto=format&n=NJMmzf2FKp5xMHlq&q=85&s=c25e11f0165f8a4638860c57c3bf82c7)
1. Confirm the measurement scope
Open Prompts → Library. Check that active Prompts reflect the questions your customers ask in the Project’s Market. Remove irrelevant questions and correct wording before you build a long-term trend. Note the tracked Competitors, because changing that set changes the Share of Voice comparison. Best practice: keep a small stable set of core Prompts across Analyses. Add new questions when your business changes, and interpret the resulting trend with that scope change in mind.2. Separate visits from response measurements
In Project Overview, read AI Visibility and its trend, then compare the same percentage with tracked Competitors without requiring Google Analytics. The AI referrals card adds identifiable visits after you connect and sync a GA4 property. It is independent of an Analysis and does not measure unique people or all AI influence. Missing setup or a first sync is not a measured zero. For your first Analysis, use Share of Voice for the response comparison, then open Dashboard → Analytics → Mentions. Under Mentions by Prompt, expand a relevant Prompt and select a Mention to read the recorded AI Response. If there is no Mention to open, check the Prompt set and Analysis History before treating the absence as a content problem.3. Compare Brands and sources
Open Dashboard → Analytics → Share of Voice with the Mentions share basis selected to compare your Brand with tracked Competitors in Mentioning Responses. Select a cell in Share of Voice by engine to inspect the answers behind it. Then review Mentions and Citations separately. In Citations Explorer, By prompt opens the question’s source links and recorded answer; By page starts from a cited URL. A Brand Mention does not require a source link, and a Citation is not the same as a Mention. See Share of Voice. Look for repeated patterns across Prompts and AI Platforms before prioritizing work. One response is an example, not a trend.4. Review Website findings separately
Open Technical Audit for the Website’s latest Audit Run. Start with Fix first and the Findings that apply to pages you control and can verify; each one says who can fix it. The Audit Run has its own timing and its own score, AI Readiness; it does not update just because an Analysis completed. When you make a fix, publish it first, then select Check if it’s fixed in the Finding to check the live pages. Run the audit again to confirm across the whole Website, and compare runs with the same start address and page limit where possible. See Check whether a Technical Audit fix worked.5. Choose one next action
Pick a supported change with clear evidence: refine an irrelevant Prompt, correct a tracked Competitor, investigate a repeated content gap, or fix a confirmed Technical Audit issue. Record the date and scope of the change. Read later results as observations rather than guaranteed effects.Related
Quickstart
Finish Project setup before reading results.
Core metrics
Understand each measurement and its denominator.