Measure AI search visibility by tracking a fixed set of buyer prompts across AI engines each month and recording how often your brand is named. Add first-party data from Bing Webmaster Tools and AI-referred traffic in your analytics, and report share of answer alongside leads, not in place of them.
Step 1: build your prompt set
Start with twenty to thirty real questions. Get them from sales calls, support tickets and search data, not from a keyword tool alone. Mix them across the journey.
- Category questions: "best [category] companies in [market]"
- Comparison questions: "[you] vs [competitor]"
- Problem questions: "how do I fix [problem]"
- Brand questions: "what is [your company] known for"
Step 2: run them consistently
Run the same prompts on the same engines every month: ChatGPT, Gemini, Perplexity, Copilot and Google AI Overviews. AI answers vary, so run each prompt more than once and record a rate rather than a yes or no.
Step 3: record what matters
For each prompt, capture four things. Were you named? Were you cited with a link? Who else was named? Was what it said about you accurate? Inaccurate statements are a finding in themselves, because they point to weak or conflicting information about your brand.
Step 4: add first-party data
Prompt sampling is an estimate, so back it with data you own.
| Source | What it tells you |
|---|---|
| Bing Webmaster Tools | Citations in Copilot and Bing AI answers |
| Google Search Console | Clicks and impressions, which include AI features in search |
| GA4 referral traffic | Visits from chatgpt.com, perplexity.ai and similar |
| Server logs | Which AI crawlers are visiting, and how often |
| Sales conversations | "How did you hear about us?" now often includes AI tools |
Step 5: report it properly
Show leadership share of answer against your top competitors, the trend over time, and a short list of fixes. Keep it next to pipeline, rather than replacing it. AI visibility is a leading signal, and leads are still the outcome.
Key takeaways
- Use real buyer questions, not only keywords
- Run prompts repeatedly and record a rate
- Accuracy of what AI says is a metric too
- Report share of answer beside pipeline