AI Visibility Checker for Real Buyer Questions
When someone asks ChatGPT, Gemini or Perplexity what to buy in your category, does your brand make the shortlist? Send a domain and the category in a buyer’s words, and we run five buyer questions across all three engines.
What the check takes
yourbrand.comcategory, in a buyer's wordsup to 3 competitorswho buys (optional)
No account, no card. We reply with the result, and we tell you plainly if we think there is nothing here worth paying us for.
Read this first
There is no instant tool on this page. This is a manual check that a person runs and reviews. We will review what you send and confirm next steps by email. We are not going to put a fake progress bar in front of a form. When a self-service checker exists, this page becomes it — the fields above, the engines and the output below are already what it will do.
What you get back
Five buyer questions in your category, each run against all three engines, with the evidence attached.
1 · The five questions we ran
Written out in full, in the buyer’s words, so you can judge whether they are the ones that matter to you.
2 · Result per question, per engine
Recommended, mentioned only, or absent — with the competitor that occupied the answer instead.
3 · Citation evidence
Which sources the answers leaned on, and whether any of them were pages you own.
4 · What we could not establish
Anything the sample was too small to support. This section is present even when it is the longest one.
- whether your brand is recommended, per engine
- which competitors are recommended instead
- how the three engines disagree with each other
- which citations and sources appear in the answers
- one high-value commercial gap worth investigating
What five questions cannot tell you
Five questions is a sample, not a survey. It will show you whether there is a visible problem; it will not describe the shape of it across your whole category, and it is not enough to build a plan on. That is what a pilot is for, and you should hear it here rather than after paying.
What the output looks like
A per-question, per-engine result with the competitor that occupied the answer — not a single score.
| Buyer question | OpenAI | Gemini | Perplexity | Who won instead |
|---|---|---|---|---|
| best dynamic QR code generator, editable destination after printing | Lost | Lost | Lost | Uniqode |
| best dynamic QR code generator, no monthly subscription | Recommended | Lost | Lost | Uniqode |
| best dynamic QR code generator, lifetime access | Recommended | Lost | Lost | Uniqode |
Disclosure
Those are real measured results from a founder-owned validation brand, not a paying customer and not customer proof. We show our own numbers because they are the only ones we can publish honestly.
How the check works
Four steps, all of them done by a person today.
You send
A domain and a category. That is the whole input — there is no form to fill in and no account to create.
We run it
Five buyer questions across ChatGPT, Gemini and Perplexity, through the official provider APIs.
A person reads it
Every check is reviewed before it is sent. During the pilot that is a founder, not a pipeline.
You get a reply
The five questions, what each engine answered, and what the sample does not support. By email.
- You send four fields. Domain, the category in a buyer’s words, up to three competitors, and optionally who buys.
- We write five buyer questions. Real purchase-stage questions in your category — “best X for Y”, “alternatives to Z” — not a generic prompt list.
- We run them across three engines. OpenAI, Gemini and Perplexity, through their official APIs with web search or grounding enabled.
- A person reads the result before you do. Including the case where we tell you there is nothing here worth paying us for.
| Field | Why we need it |
|---|---|
| Domain | To detect citations to pages you own, and to separate your own properties from competitors. |
| Category, in a buyer’s words | The questions are built from this. “Dynamic QR code generator” produces a very different set than “marketing platform”. |
| Up to 3 competitors | Displacement questions — “alternatives to X” — are usually where the sharpest findings are. We also report competitors we were not told about. |
| Who buys (optional) | Lets us add segment-qualified questions, which find gaps that a generic “best X” hides. |
What an AI visibility checker actually is
Different tools mean very different things by “checker”. For buyer visibility the useful question is narrow: what does the assistant recommend when a buyer asks for a product like yours?
AI readiness checker
Audits technical and content signals on your website — crawlability, structure, files.
AI visibility checker
Measures the answer layer itself: whether AI systems mention, recommend or cite you. A site can be perfectly accessible to AI crawlers and still lose every commercial recommendation question, because being retrievable and being chosen are different problems.
Mentioned≠Recommended
If the answer says “Brand A is the best fit. Brand B also exists but is less suited to this use case,” both brands are visible and only one wins. Those two states imply completely different work, so they are never merged.
Which questions to test
The best first test is not your brand name. It is a category question a buyer could ask without ever having heard of you.
B2B SaaS
“What are the best [category] tools?” · “Which [category] platform is best for [ICP]?” · “What are the best alternatives to [leader]?” · “Which [category] offers [differentiator]?”
Local service business
“Who are the best [service] providers in [city]?” · “Which [service] company is best for [specific need]?”
Publisher
Less about vendor recommendation: which sources are cited, whether your site is used as supporting evidence, and which information questions competitors own.
If the answer recommends three competitors and not you, you have found a real decision gap.
What happens if the evidence is weak
The check does not manufacture a diagnosis.
The honest result
If there is one unstable observation or too little compatible evidence, the result is “more evidence required” — not “rewrite your pricing page”. That is a core trust rule, and it is the reason a check can come back saying we found nothing worth acting on.
Two tools running “the same” check can return different results because of the model or provider, the time, the location, web-grounding behaviour, prompt wording, hidden context and how many times the question was repeated. That does not make measurement useless. It means the protocol has to be visible — ours is below.
What a full audit adds
A quick check can reveal a gap. A pilot turns it into a plan you can defend.
Approved question cohort
A stable, versioned set that can be measured repeatedly and compared later.
Competitor normalisation
Product competitors separated from publisher and citation-source domains.
Strategic positioning
The claims you actually want the market to associate with you, tracked per engine.
Owned-site inspection
A bounded read of the public pages that may support the buyer decision.
ProductTruth
Whether a capability is verified true, verified false or unknown — before any content change is recommended.
Remeasurement
The same question measured again after the work, under a compatible protocol.
What to do with a failed result
Do not respond to every loss by creating another blog post. Work the evidence in order.
- Confirm the pattern with enough compatible observations.
- Inspect the answer and its citations.
- Identify the evidence themes that keep appearing in the answers that won.
- Inspect the relevant page you own.
- Confirm the product fact before you publish a claim about it.
- Decide whether the gap is actually addressable.
- Make one defensible change.
- Remeasure under a compatible protocol.
Common questions
Is an AI visibility checker the same as an AI SEO audit?
No. A visibility checker measures the generated answers themselves. An AI readiness or SEO audit inspects your website. A site can pass a readiness audit and still lose every commercial recommendation question, which is why the two are not substitutes.
Can I check ChatGPT visibility manually?
Yes — open an assistant and ask a buying question in your category. What you cannot do by hand is run the same questions repeatedly across engines, keep the raw answers, exclude failed calls from the denominators and compare against a frozen baseline later.
Why not just search my brand name?
A branded question tells you whether the assistant knows you exist. Category and use-case questions tell you whether it recommends you to a buyer who has not already chosen you — which is the version with revenue attached.
Does one check prove my AI ranking?
No. AI answers vary between runs, models, locations and dates. One check is a snapshot that can reveal an obvious problem; it is not a ranking and not enough to build a plan on.
What should I do if a competitor is recommended instead?
Confirm the pattern with enough compatible observations, then inspect the answer evidence, the cited sources, the relevant page you own and the verified product facts before choosing an action. Responding to every loss with another blog post is how budgets disappear.
How we measure, and how this page is sourced
What we actually measure
Every result comes from the official OpenAI, Google Gemini
and Perplexity APIs with web search or grounding enabled. In our data those
surfaces are recorded as openai_api_web_search,
gemini_api_google_search and perplexity_api_sonar, and reports name
them explicitly.
That is not the same thing as a logged-in person using the ChatGPT app. These are official interfaces with web search on, but they are measurement surfaces rather than consumer sessions: no personalisation, no memory, no app-only features. Results can differ from what any individual user sees, and we do not claim the two are identical. We would rather put that on the pricing page than in the small print.
How this page is sourced
- Vendor pricing and features come from each vendor's own public pricing page, with the date we checked it. We do not copy figures out of other people's comparison posts.
- Google-specific claims come from Google Search Central, not from SEO commentary about Google.
- Research claims name the paper and separate what was demonstrated in a benchmark from what holds in the open web.
- First-hand numbers come from our own measurement runs. Where the brand measured is founder-owned, the page says so in the same sentence as the finding.
This page is maintained by the person who built the measurement engine. There is no editorial team and no external reviewer behind it, and we are not going to invent one for a byline. How the measurements are collected.