Free tool · Comparison

Who’s winning instead of you?

Ask any AI engine for the best plumber, dentist, or inn in your town and you don’t get a pass or a fail. You get a list with three to five names on it. If you’re not on that list, it isn’t because the engine forgot you — it’s because somebody else is standing in the slot. You already know you’re invisible. The question worth ninety seconds is who’s in the slot, and what have they got that you haven’t?

Every visibility score measures one business against itself

There is a real industry now in telling you whether ChatGPT mentions your business. We build several of those tools ourselves and they are useful. But every one of them — ours included, until this one — measures a single business in isolation. Present or absent. Sixty-one out of a hundred. Better or worse than last month.

That framing quietly hides the thing that decides the outcome. Local recommendation answers are competitive and the field is small: three names, sometimes five. Every slot you don’t occupy is occupied by somebody in your town who did something you didn’t. That “something” is short, specific, and almost always visible from the outside if you go and look.

Why not just report a percentage?

Because a percentage is a scoreboard, and a scoreboard is not an instruction.

Telling a one-van plumber he owns twelve per cent of his town’s AI visibility tells him he is losing. He knew that. It does not tell him what to do on Tuesday evening. And a percentage of that kind depends entirely on which questions were asked, how often, and of which engines — the query set, the sampling cadence, the engine list. When those aren’t published, the number is a lie with a decimal point in it.

So this tool doesn’t report one. It reports the difference between you and the businesses in the slots, in the order worth fixing.

The eight signals

Two rules decided what got on this list, and both are strict.

• Engines actually weigh itFor local recommendation queries specifically — not for general web ranking, which is a different problem with a different answer.
• You can verify it in about a minuteOn a competitor, from public pages, with no SEO knowledge and no tooling. Things an engine also weighs that you cannot see from outside — its training data, its retrieval index, its own quirks — are deliberately absent. A checklist you can’t check is decoration.

What survived: a page for that exact service in that exact town; plain-text answers to the questions customers actually ask; review count and recency; appearing in somebody else’s “best of” list; a real named human on the About page; published prices or ranges; hours and contact details that agree everywhere you look; and independent press coverage. They are weighted, and the weighting is our editorial judgement, said plainly on the tool itself rather than dressed up as a measurement.

The gap ledger

Mark who has what — you, then each competitor — and the output builds itself in three blocks.

• Do these firstEvery competitor you checked has it. You don’t. When the difference between you and the businesses in the slot is unanimous, it stops being a theory and starts being the job.
• Then theseSome of them have it and you don’t. Real, but less certain it’s the deciding factor. Finish the block above first.
• Say this louderYou have it and none of them do. Nobody else prints this block, and it is often the most valuable one. Most owners are sitting on two or three real advantages they have never written down in plain words on their own website — which is precisely where an engine would have to read them to repeat them.

What this doesn’t do, said plainly

It does not query the engines for you. You run the question, read the names, type them in. Automated multi-engine querying costs real money per lookup, which is why the platforms doing it start around thirty-nine dollars a month on metered credits and climb from there — and why one of the better ones bundles the credit pool with its rank tracker, so an AI check competes for quota with a map scan and independent reviewers call the billing hard to forecast.

It also doesn’t track you over time. One run is a sample. Run the same question a week later, compare the names, and only then trust the ordering.

What it does do, for nothing, is the part that turns an observation into an instruction.

Questions

The other six

All seven tools share one set of business details — fill them in once and switch between tools without retyping anything.