Start Here

Check your own visibility, free

You can find out whether Google, Google Maps, and AI assistants like ChatGPT actually recommend your business – today, yourself, without buying anything. This page is the complete method: the test, the pitfalls that quietly ruin it, and how the platforms differ. Nothing is held back. What a page can’t give you is experienced judgment about what your results mean – more on that at the end, honestly.

Before you start: three ground rules

Test signed out. Search engines and AI assistants personalize. If you run the test logged in to your own accounts, on the machine where you visit your own website daily, the results will flatter you. Use a private browsing window, signed out, and you’ll see something closer to what a stranger sees.

Ask customer questions, not marketer questions. Your customers don’t type “plumber Concord NH best rated.” They type “my water heater is leaking, who do I call in Concord.” Test the question that actually gets asked, or you’re measuring the wrong thing.

Date everything. Write your results down with the date. AI answers change; a dated baseline is the only way you’ll ever know whether anything you fixed actually moved.

The manual visibility test

  1. Write down the questions your customers actually ask

    Five to eight questions, in the words a real customer would use – full sentences, plain language, your town’s name the way people say it. Cover both moods: the emergency (“my water heater is leaking, who do I call in [your town]”) and the considered (“who’s the most trustworthy [your trade] in [your town]”). Not keyword strings. The phrasing is most of the test.

  2. Ask them everywhere your customers ask them

    Run every question in all four places: a regular Google search, Google Maps, ChatGPT, and at least one more AI assistant. Same wording in each. One platform is not a proxy for the others – the section below on how the engines differ is why.

  3. Ask more than once

    AI assistants don’t give the same answer every time. Run each question at least twice, ideally on different days. One appearance is luck; a pattern is visibility. This single habit separates a real read from a false one.

  4. Write down exactly who gets named

    For every question, on every platform, record which businesses appear and in what order – including whether you appear at all, and including every business you’ve never heard of. Date the notes. If the AI names three competitors and not you, that’s the finding; write it down straight.

  5. Check what they say about you

    Now ask each platform about your business by name: what you do, where you work, whether you’re any good. AI assistants get facts wrong about small businesses constantly – wrong services, wrong towns, stale hours, a competitor’s phone number. Every error is a finding, and most of them are fixable.

  6. Fix in priority order

    Start with your Google Business Profile: complete, accurate, right categories, real photos, reviews you actually answer. It feeds Maps directly and influences AI answers. Then your website: a plain page for each service, a page that says who you are and where you work in ordinary sentences, and a real FAQ written the way this page’s questions are written. This site is built exactly that way, on purpose – use it as the reference.

  7. Re-run the same test and compare

    A month or two after the fixes, run the same questions – same wording, same platforms, that’s the point – and compare against your dated baseline. Movement you can see beats any promise anyone can make you.

The engines are not interchangeable

The single most common mistake is treating “AI” as one thing. It isn’t. Each platform builds its answer differently, and a business can look fine on one and be invisible on another – at the same time, in the same town.

Roughly, as of today: Google’s AI answers sit on top of Google’s own search index and Google Business Profiles, so they inherit whatever your Google presence already is – strengths and problems both. Chat assistants like ChatGPT build answers from a mix of what their models learned over time and what they pull from the live web when asked, which means they lean heavily on what other sources say about you – reviews, directories, mentions, articles – not just your own site. Assistants that cite their sources tend to reward pages that answer a question plainly in text, because those are the pages worth citing. And every one of them refreshes on its own schedule, so a fix can show up on one platform weeks before another.

Two practical consequences. First, you have to test all of them – a great Maps presence tells you nothing about what ChatGPT says. Second, “why am I invisible” has a different answer per platform: missing on Google’s AI usually points at your Google footprint; missing in a chat assistant usually points at what the wider web says, or doesn’t say, about you. Same symptom, different disease, different fix.

Six ways this test quietly goes wrong

  • Testing logged in

    Your own history flatters you. Every result you collect while signed in is contaminated, and the test will tell you you’re more visible than you are. Signed out, private window, every time.

  • Asking the marketer’s version of the question

    “Best plumber Concord NH 2026” is not a question a customer asks, and the answer to it isn’t the answer your customers get. If the phrasing wouldn’t come out of a real person’s mouth on a bad day, don’t test it.

  • Testing once and believing it

    One run is an anecdote. Answers vary between runs, days, and phrasings – that’s normal, not broken. A conclusion drawn from a single answer, good or bad, is the most common false read in this whole exercise.

  • Reading every absence as the same problem

    “I’m not showing up” has at least four different causes: your business data is thin or wrong, the web says almost nothing about you, your site doesn’t answer questions in plain text, or your reputation signals are weak. Each has a different fix. Treating them as one problem is how people fix the wrong thing first.

  • Buying a shortcut

    Anyone promising guaranteed AI placement, “one weird trick” optimization, or instant citations is selling something nobody can deliver. There is no submit-your-site-to-ChatGPT form. The unglamorous work above is the work.

  • Fixing everything at once

    If you change ten things and something improves, you’ve learned nothing about what worked – and you can’t repeat it. Fix in priority order, re-test, then move down the list. Slower, and the only version that teaches you anything.

What the test can’t tell you

Run honestly, the test above gives you a real snapshot: where you appear, where you don’t, what’s wrong in what the AI says about you. That’s genuinely valuable, and it’s yours for an afternoon’s work.

What a snapshot can’t give you is the story. Is that missing answer noise or signal? Why is this platform naming you and that one not – and which of the four causes of absence is yours? Which fix moves which engine, and in what order, so you’re not spending months on the wrong thing? Those calls come from having run this across many businesses and watched what actually moved – pattern recognition you can’t get from your own single data point, however carefully you collect it.

That’s what the free audit is: your snapshot, read by experienced eyes, sent back in plain language – what’s working, what’s broken, what to fix first and why. It’s free and there’s no obligation, so the honest math is simple: run the test yourself, absolutely, and then lose nothing by having someone who does this every day tell you what your results mean.

Get your free visibility audit

Prefer to watch it taught?

The Get Found YouTube channel teaches this material in video, step by step, in order. Same method, same no-hype rules – use whichever format suits you.

Watch Get Found on YouTube