How to measure your brand's visibility in AI search
A practical framework for measuring AI search visibility: build a question set, track mentions, citations, recommendations and sentiment, and benchmark competitors.
You can't improve what you don't measure, and AI search is harder to measure than traditional rankings. There's no single position number, answers vary between sessions, and most platforms don't share analytics. A consistent, repeatable method still gives you a useful picture.
1. Build a fixed set of questions
Write down 20–50 questions your potential customers would realistically ask. Mix in different kinds of questions so you see the full picture:
- Service: "Who offers emergency boiler repair?"
- Local: "Best dentist in [city]"
- Informational: "How much does a roof replacement cost?"
- Comparison: "[Your brand] vs [competitor]"
- Branded: "What is [your brand] known for?"
2. Pick the platforms that matter
Most businesses should at least cover ChatGPT, Google AI Overviews, Gemini and Perplexity. Run the same questions on each one, because they draw on different sources and often give different answers.
3. Record the right signals
- Mention: is your brand named at all?
- Citation: is your website linked as a source?
- Recommendation: are you actively suggested as a good option?
- Position: where in the answer do you appear?
- Sentiment: is the description positive, neutral or negative?
- Accuracy: are the facts about you correct?
4. Benchmark competitors
Note which competitors appear for each question. Share of voice, meaning how often you're mentioned compared with competitors across your question set, is often more useful than looking at your own mentions alone. It shows who AI recommends instead of you and for which kinds of questions.
5. Calculate a simple score
Turn your results into a score you can track, such as the percentage of questions where you're mentioned, split by platform and by question type. Keep the method identical each time so changes reflect real movement rather than a different approach.
6. Repeat on a schedule
AI answers change as models and sources update. Re-run the same set monthly or quarterly. Single snapshots can be noisy, so look at trends.
Doing it manually vs. with a tool
A spreadsheet works, but running dozens of questions across four platforms and scoring every answer takes hours. The AI Visibility report automates this: it generates relevant questions, runs them across the major platforms and reports your score, mentions, citations, competitor share of voice, coverage by question type and sentiment.
Example: a simple AI visibility log
Run the same prompts each month and record the result. Made-up example for a dentist:
| Prompt | Tool | Mentioned? | Cited source | Competitor named |
|---|---|---|---|---|
| best family dentist in Boise | ChatGPT | No | — | Smile Co. |
| best family dentist in Boise | Perplexity | Yes (#3) | Yelp listing | Smile Co. |
| emergency dentist open Saturday Boise | Gemini | Yes (#1) | Own website | — |
Frequently asked questions
Can I see AI traffic in Google Analytics?
Some AI platforms send referral traffic that shows up as referrals, but many AI answers don't produce a click, so analytics only shows part of the picture.
How many questions should I track?
Enough to cover your main services and question types. Twenty to fifty is a practical range for most small businesses.
Why do I get different answers each time?
AI responses can vary from session to session. That's why fixed question sets and trends matter more than any single answer.
Does Search Console report AI Overviews?
AI Overview appearances are included in overall Search performance data, but they aren't currently broken out separately.
How often should I measure?
Monthly or quarterly works for most businesses, plus a check after any major changes.