Do You Appear in AI? How to Measure Your Visibility
8 min · 2026-06-26
Most business owners don't know the answer to "is AI recommending me?" They guess. But visibility isn't something to guess — it's something to measure, just like you measure your Google ranking. In this article I explain how to systematically measure your visibility in ChatGPT, Gemini and Perplexity, which questions to ask, and how it all condenses into a single score (AIVI). Everything you do without measuring is shooting in the dark.
- AI visibility is measurable; you measure it with a fixed method instead of "feeling" it.
- Proper measurement needs a fixed question set + multiple engines + regular repetition; a single query misleads.
- Measurement looks at three things: can AI find you, does it recognize you correctly, does it cite you.
- AIVI reduces these three layers to a single visibility score; the real value is in the before/after comparison.
Why "feeling" doesn't work
When I ask a business owner "do you appear in AI?", I usually get one of two answers: "I think so" or "I don't know". Both lead to the same place — no data. Seeing your name once in ChatGPT doesn't mean you "appear"; in another question, another engine, another day, you may vanish completely. AI answers are variable. That's why you need systematic measurement, not a one-time glance.
Without measurement there is no improvement. If you don't know where you're invisible, you can't know what to fix. The goal of this article is to turn that "darkness" into a measurable table.
The 3 rules of proper measurement
1. Use a fixed question set
Ask the same questions every time. If you change the questions each round, you can't compare. The set should cover three types: questions in the customer's problem language ("AI doesn't see my site, what do I do"), category questions (without your name, "who is the best ..."), and brand questions (with your name). These three types probe different layers of visibility.
2. Test across multiple engines
Appearing in ChatGPT doesn't mean you'll appear in Gemini. Each engine has its own data source and behavior. For a realistic picture, measure at least in ChatGPT, Gemini and Perplexity; ideally Claude and Bing/Copilot too. Ask the same question to each engine separately and record the result.
3. Repeat regularly and document
Visibility isn't a one-time photo, it's a curve. Today's "zero" measurement (the baseline) is your reference point. As you produce content and strengthen signals, you re-ask the same questions and watch the change. Document every result with a screenshot — this is both proof of your progress and concrete evidence that your work is paying off.
How a visibility score is calculated
To turn raw observations into a comparable number, a simple scoring works. For each question-engine combination, assign a score:
| Score | Meaning | Example |
|---|---|---|
| 0 | You're absent | Your name isn't mentioned, or the wrong person/brand appears |
| 1 | Weak / indirect | Your name appears but with wrong info, or lumped in with competitors |
| 2 | Clear and strong | You're mentioned with correct info, or your site is cited as a source |
Add up all scores and divide by the maximum possible. For example 6 questions × 5 engines = 30 measurements, max 60 points. If you scored 12, your visibility score is 20%. That number alone doesn't matter; what matters is repeating the same measurement a month later and seeing it rise from 20% to 35%. That's what AIVI does.
AIVI: reducing three layers to one score
AIVI (AI Visibility Index) puts the logic above into a standard method. It measures visibility in three layers: can AI find you (visibility), does it recognize and trust you correctly (knowledge and trust), how often it cites you (citation). Combined, they produce a single score. The value isn't in one photo but in the curve over time: measure, close the missing signals, measure again, show the difference.
- To start yourself: write a fixed set of 5-6 questions, ask them in three engines, log them in a table with the 0-1-2 scoring from this article.
- Save the results with screenshots; this becomes your "Month 0" baseline.
- A month later, ask the same set again; see the difference.
Frequently asked questions
Can I measure my AI visibility myself?
Yes, you can do a basic measurement yourself: write a fixed question set, ask it in ChatGPT, Gemini and Perplexity, and log the results with a simple 0-1-2 score in a table. This gives you a realistic starting snapshot. Deeper, consistent measurement (multiple engines, regular repetition, layer weights) usually requires expertise and a system.
How many engines should I measure in?
At least three: ChatGPT, Gemini and Perplexity. Ideally add Claude and Bing/Copilot too. Measuring in a single engine is misleading because each has a different data source and behavior; you might appear in one and be completely invisible in another.
How often should I measure?
Monthly is a good start. If you're producing content or actively working on visibility, measuring every two weeks shows the change more clearly. The key is using the same question set and same engines every time; only then is the before/after comparison fair.
What if my AIVI score is low?
A low baseline score is normal, especially for a new site. What matters isn't the score itself but seeing the right steps to raise it: entity clarity (structured data that tells AI who you are), technical accessibility, citable content, and external citations. Measurement shows you which layer you're weak in — that is, where to push.