How to measure your brand's visibility in AI answers
A clear, repeatable method for measuring how often ChatGPT, Gemini, Claude and Google AI name your brand, and how to tell whether your work is paying off.
To measure AI visibility, choose a fixed set of real buyer questions, ask each one across several AI engines multiple times, and record whether you're named, in what position, how accurately, and which sources are cited. The share of answers that name you is your Share of Answer. Repeat the same test on a regular schedule to track progress.
What to measure
| Metric | What it tells you | How to record it |
|---|---|---|
| Share of Answer | How often AI names you | Answers naming you ÷ total answers |
| Position | Whether you're named first or last | Rank within each answer |
| Accuracy | Whether AI describes you correctly | Correct, partly correct or wrong |
| Citations | Which sources AI relies on | Domains and URLs cited |
| Competitors | Who wins when you don't | Brands named per question |
Step 1: Build your question set
Start with 20 to 50 questions buyers really ask, phrased the way they'd ask AI. Mix:
- Category questions: "Best employment lawyer for small businesses?"
- Problem questions: "How do I fix a leaking flat roof?"
- Comparison questions: "Alternative to the market leader for small teams?"
- Brand questions: "Is [your brand] good for X?"
Our list of questions buyers ask AI can help you get started.
Step 2: Choose engines and repeat runs
Test at least ChatGPT, Gemini, Claude and Google AI. Ask each question several times, because answers vary between runs. Five runs per question per engine is a solid baseline. Use clean sessions without personal history.
Step 3: Record results consistently
Use the same spreadsheet or tool every time, with the metrics in the table above. Consistency matters more than perfection: the value is in comparing like with like over time.
Step 4: Calculate Share of Answer
Divide the number of answers naming you by the total number of answers. For example, 20 questions × 4 engines × 5 runs = 400 answers. If 60 name you, your Share of Answer is 15%. Break it down by engine, service and market to see where to focus. See Share of Answer explained.
Step 5: Turn data into actions
- Questions where competitors win show where content or mentions are missing.
- Inaccurate answers point to inconsistent facts somewhere online.
- Frequently cited sources show where to earn mentions.
This is the core of a TUNE Audit.
Step 6: Report on a fixed schedule
Monthly reporting suits most businesses. Keep the question set stable and add new questions as a separate group, so trends stay comparable. The Resonance Retainer includes monthly reporting, and our case studies show what before-and-after results look like.
Tools vs doing it by hand
Manual testing is fine to start. As your question set grows, tools save time. We compare options in AI visibility tools.
- Use a fixed set of real buyer questions.
- Test several engines, several times each.
- Share of Answer is the headline metric; accuracy and citations explain it.
- Report monthly on the same questions to see trends.
Frequently asked questions
How many questions should I track?
Start with 20 to 50 covering your main services, then expand.
Why repeat each question?
AI answers vary from run to run. Repeats smooth out the noise so real changes stand out.
Should I track brand-name questions?
Yes, separately. They show accuracy, while unbranded questions show whether you're recommended.
How often should I measure?
Monthly for most businesses; weekly if you're making rapid changes.
Explore next
Is AI recommending you, or your competitors?
Find out free in two minutes.