What does the AI Visibility Checker do?

It asks a group of AI engines whether they can find and recommend your business. You paste your URL and add your business name, category and city, and it generates buyer-intent prompts, runs them across ChatGPT, Gemini, Claude, Perplexity, Groq, Mistral and Llama, then returns a website score, an AI score and an overall grade.

Do I need to fill in business name, category and city?

You should. The checker uses them to build realistic prompts that match how a buyer would actually ask about your business. The more accurate those fields are, the more useful your AI score is.

What do the pass, warn and fail checks mean?

Each check is a concrete signal — indexability, structured data, content depth, trust signals and more. Green is where you are solid, amber is worth reviewing, and red is what is most likely holding back your AI visibility.

How is the AI visibility score different from the website score?

The website score measures the technical and content readiness of the site itself. The AI score measures whether the AI engines actually surface and cite you. A site can have a strong website score yet still be invisible in AI answers, which is exactly what the gap between the two reveals.

AI Visibility Checker: The Complete How-To Guide

Most people find the AI Visibility Checker, paste a URL, stare at the score, and then have no idea what to do with it. That's a shame, because the tool gives you a lot more than a number — it gives you a list of the exact things standing between you and being recommended by ChatGPT, Perplexity and Gemini.

This guide walks through it the way I'd walk a client through it. Paste the URL, but then go a bit deeper.

What the tool is actually doing

Under the hood it simulates a buyer. It takes your URL, your business name, your category and your city, and it writes the kind of prompts a real customer would type — questions like "best [category] in [city]" or "who is [business name] and are they legit." Then it runs those prompts across seven engines: ChatGPT, Gemini, Claude, Perplexity, Groq, Mistral and Llama.

For each engine it checks three things. Did it mention you at all? If so, did it match you on your business name or your domain? And what did it say — is it praising you, describing you, or ignoring you? That engine-by-engine breakdown is the part most people skip, and it's the most useful part.

Filling in the fields properly

The URL is the only required field, but you're leaving signal on the table if you skip the rest.

Business name — Use the exact name your customers search for. If your legal name and trading name differ, use the trading name. It's the one the AI will recognise.
Category — Be specific. "Plumber" beats "home services." The more precise, the more realistic the prompts the checker generates.
City — This matters hugely for local businesses. AI answers are increasingly localised, and the checker uses it to build location-aware prompts.

The tool auto-generates the buyer-intent prompts, competitor comparisons and recommendation checks from your profile. There's no manual mode to fumble — you just fill in the fields accurately and let it run.

Reading the three scores

The result screen gives you a website score, an AI score, and an overall grade. They're answering different questions, and the gap between the first two is where the interesting insight lives.

Website score — How ready is the site itself? This covers indexability, structured data, content depth, security, trust signals and readability. It's the foundation.
AI score — Does the AI actually surface and cite you? This is the part you can't see in Google Search Console.
Overall grade — The combined picture, an A–F style grade at a glance.

If the website score is high and the AI score is low, you have a classic attribution problem: your site is technically healthy but the AI engines aren't choosing it as their source. That's fixable, and the checks panel tells you where.

Understanding the pass / warn / fail checks

Below the score you'll get a list of checks, each marked pass, warn or fail. This is the part of the tool I'd personally work from.

Pass means that signal is solid. Leave it alone.
Warn means it's present but weak. Worth reviewing, but not urgent.
Fail means it's the most likely reason AI engines don't trust or cite you.

Don't try to fix everything at once. Fix the fails first — they're the ones actively costing you — then circle back to the warns. Rerun the check after each round of changes and watch each item flip from red to amber to green.

Reading the engine-by-engine results

A brand can be beloved by ChatGPT and invisible to Perplexity. That's normal, and it's why the per-engine view matters. If you're strong on one engine and not another, look at what differs — sometimes it's just that one engine leans on different sources, sometimes it's that your information is inconsistent and one engine found the coherent version.

Look at how each engine matched you: on your name, on your domain, or not at all. Domains are a stronger signal than a bare name mention, because a domain match means the AI trusts an actual page of yours rather than a directory listing.

What to do after you have the report

The checker is a starting point, not the finish line. Once you know your baseline, the workflow is:

Fix the failed checks on your site.
Make your key facts consistent and explicit — business name, address, category, services, history — so an AI can confidently describe you.
Add the structured data the report flags as missing.
Rerun the check on a schedule, not once.

If you want this done for you rather than doing it yourself, that's literally the service we run at Launch at Dawn. Book a free website teardown and we'll walk you through your own report, engine by engine.