Enter your brand and what you sell, plus your market, audience and competitors if you like, and this tool writes 5 to 9 buyer prompts to run in ChatGPT, Gemini, Perplexity and Google AI. Record what each answer did and it computes your mention rate and a verdict of Invisible, Emerging or Known.
Step 1 · Your brand
Step 2 · Ask the engines
Paste each prompt into a fresh chat (a temporary or logged-out chat is best). For every engine, pick what happened. Leave a cell on "Not tested" if you skip it.
| Buyer prompt | ChatGPT | Gemini | Perplexity | Google AI |
|---|
Brand knowledge (not scored)
Ask these too and check the facts AI gives about you. Wrong or outdated facts are a separate problem from not being mentioned.
The tool above gives you a quick, honest read on whether AI engines name your brand when buyers ask about your category. It does not query any AI itself: it writes the prompts, you run them, and it scores what you record. That keeps it private and lets you see every answer with your own eyes.
How do you use it?
- Enter your details. Brand name and what you sell in buyer words (for example “trail running shoes” or “vegan bakery”) are required. Website, market or city, who it is for, and up to three competitors are optional.
- Generate your prompts. The tool writes five core buyer prompts, such as “What are the best vegan bakery brands in Lisbon?” or “Which trail running shoes would you recommend for beginners?” (illustrations). Competitors add up to four more: an “alternatives to” prompt for each competitor, and a head-to-head prompt for the first. That gives 5 to 9 prompts.
- Run each prompt in four engines. Copy it into ChatGPT, Gemini, Perplexity and Google’s AI answers (AI Overviews or AI Mode), ideally in a fresh, temporary chat.
- Record each result. Every prompt and engine has its own cell. Choose Not mentioned, Mentioned, Top 3, or Top 3 + linked to you (named in the first three and linked to your site). Leave a cell on Not tested if you skip it.
- Ask the three brand knowledge prompts. “What is [your brand]?”, whether it is any good, and who it is for and what it costs. These are not scored. They show whether AI has your facts right, which is a separate problem from not being mentioned.
- Score it. With at least four results recorded, the tool shows your mention rate, a verdict, how many answers put you in the top three and linked you, and a count per engine.
How is the verdict worked out?
The mention rate is simple: answers that named you (Mentioned, Top 3 or Top 3 + linked) divided by all the answers you tested, rounded to a whole percent. Untested cells do not count either way.
| Mention rate | Verdict | What it suggests | Next step shown |
|---|---|---|---|
| Under 10% | Invisible | Too little on the web ties your brand to this topic | Getting known through mentions in real publications, in context |
| 10% to 49% | Emerging | AI knows you some of the time, not consistently | More mentions in context and consistent facts across the web |
| 50% or more | Known | AI already names you in most answers | Position: being recommended near the top, which is step 2, The Blue Ocean GPT |
Position (Top 3, linked) is reported alongside the verdict but does not change it. A brand named last in every answer and a brand named first in every answer get the same mention rate, which is why the top 3 and link counts sit next to it.
Why does it use several prompts and engines?
Single answers are unreliable. In a SparkToro and Gumshoe study (one co-author works at an AI tracking startup), there was less than a 1 in 100 chance that ChatGPT or Google’s AI gave the same brand list in two runs of the same prompt. Engines also retrieve from different sources, so a brand can be named by one and missing from another. A rate across many prompts and engines is a fairer picture of your AI visibility than any single answer.
What can it not tell you?
It reflects only the prompts you ran, once each, on the day you ran them, with your own settings. Five to eight prompts in four engines is at most 32 answers, a small sample, so a shift of a few answers can move the verdict. It cannot see why an engine left you out, and it does not judge whether the facts AI gives about you are correct; that is what the brand knowledge prompts are for.
For a sturdier measure, the guide on how to measure your brand’s AI visibility shows how to build a larger prompt set with repeated runs, and our methodology explains how we measure before and after placing mentions.
What happens to the data you enter?
The tool runs entirely in your browser, makes no calls to any server and stores nothing. Your entries and scores disappear when you close or reload the page. The Copy buttons only put a prompt on your clipboard. If you click the button under your verdict, the link to our homepage form carries your website (or brand name, if you left the website blank) and your category, so the form can start with them. Nothing is sent unless you click.
What to do next
- If your verdict is Invisible, read why AI doesn’t mention your brand.
- Keep your prompts and re-run them in a month, with the same setup, to see the trend.
- Get a free AI visibility check and we will ask the engines a wider set of buyer questions for your brand and show you where you stand.
Frequently asked questions
Is the AI visibility self-check really free and private?
Yes. The tool runs in your browser, with no account and no email. It does not send what you type to us or anyone else, and nothing is saved. You paste the prompts into the AI tools yourself and record what you see. Only if you click the button under your verdict do your website (or brand name) and category travel in the link to our homepage form, so you do not have to retype them.
Why do I get a different answer when I ask the same question again?
Because AI answers vary from run to run. A SparkToro study found ChatGPT and Google's AI almost never returned the same list of brands twice for the same prompt. That is why the tool scores a set of prompts across four engines rather than one answer. For a steadier picture, repeat the whole check on another day with the same setup and compare the two rates.
What do Invisible, Emerging and Known mean?
Your mention rate is the share of tested answers that named your brand at all. Under 10% is Invisible, 10% to 49% is Emerging, and 50% or more is Known. Cells left on Not tested are ignored, and you need at least four results for a score. The verdict covers only the prompts and engines you scored, on that day, so treat it as a snapshot.
Should I be logged in to ChatGPT when I run the prompts?
Use a temporary chat, or a logged-out window where the engine allows it, with memory and custom instructions off. Personalization and earlier conversations can push an assistant toward brands you have already asked about, which flatters your result. Use the same setup each time you re-check, so changes come from the engines and not from your settings.
My verdict is Invisible. What should I do first?
Check the basics on your own site first, such as whether your robots.txt blocks OAI-SearchBot or PerplexityBot, and whether your brand facts match everywhere. Then look at how many independent publications mention you next to your category. For most brands with near-zero AI visibility, getting known is step 1, and the guide on fixing zero visibility sets out the order.
See what AI says about your brand right now.
Tell us your website, your topic and your market in about a minute. We run a free AI visibility check and show you where you stand.
Get my free AI visibility check →