Check your AI visibility

How to check if Claude recommends your business

By the Peekl team · Last reviewed: September 27, 2026

Part of our series on Measuring your AI visibility. The reference guide is the guide to AI visibility metrics.

To check if Claude recommends your business, open Claude and ask the questions your customers ask, like who to hire near them for what you do, then see whether it names you. If you would rather have it counted for you, the Claude visibility checker puts fifteen customer questions to Claude three times each inside the Peekl audit. Claude tends to be more careful than other assistants: it may give fewer names, or ask where you are before it answers, and that caution changes what it takes to be one of the names it does give.

This guide walks through a five-step Claude check you can run in an hour, a Yes or Not yet self-score built around the way Claude decides, and what we saw when we put 45 Claude answers about one clinic next to the same questions on ChatGPT and Gemini.

Illustration of checking whether Claude recommends a local business: an AI answer panel with a sparkle, a check mark and a magnifying glass inspecting a result.

Free, no card needed.

Key takeaways

Ask Claude the questions your customers ask, with your trade and town, and note whether it names you, names a rival, or holds back.
When it has web search turned on, Claude looks things up rather than guessing, then tends to name only businesses it can describe with confidence.
Because Claude is cautious, its lists are often shorter, so the bar to be included is higher, not lower.
Repeat the question. Claude named one clinic in 13 of 45 answers, yet just one question out of 15 produced its name on every run.

How to check what Claude says about you

Open Claude and ask it the questions people ask when they are close to choosing, using your real trade and town. Read the answer for whether your name is there, and also for how the answer is shaped, because Claude often behaves a little differently from other assistants.

It may name only two or three businesses rather than a long list, note that it cannot be certain, or ask for your location before it commits. None of that is a fault. It tells you Claude sets a high bar for naming a specific business, which is exactly the bar worth understanding.

“Who should I hire for [the job you do] in [your town]?”
“Recommend a trustworthy [your trade] in [your town]”
“Best [your trade] in [your town], and why”
“Who are the most reputable [your trade] near [your area]?”
“I am in [your town] and need [a specific service] soon. Who would you go to?”

The five-step Claude check

A single question shows you Claude's manner. Five questions, asked three times each with web search on, show you whether it actually knows you. Here is the routine, tuned to the way Claude behaves.

Step 1: pick five questions from the list above and write your town into each one. Claude will ask where you are if you leave it out, and the extra turn makes runs harder to compare.
Step 2: open a new chat and make sure web search is switched on in Claude's settings (Anthropic's help page shows where the toggle lives). Without it, Claude answers from what it learned in training, which is older, and it will usually tell you it cannot verify current businesses. Do not mention your own company in the question.
Step 3: score each answer in four columns, not three: you were named, a competitor was named, Claude gave no names, or Claude declined and explained why. That fourth column is common with Claude and it is information: it means nobody in your trade and town is described clearly enough for it to commit.
Step 4: read the citations. Claude marks the pages it drew on, so you do not need to ask. Note which ones describe you and which describe the competitor it preferred. The gap between those two lists is your to-do list.
Step 5: repeat each question three times, then run the same five on ChatGPT and Gemini. If Claude names you and the others do not, you are being described clearly somewhere Claude trusts but are missing from directories and Google's record. If the reverse, your details are out there but not stated plainly enough for a cautious reader.

Score yourself: Yes or Not yet

Claude rewards clarity, so the checklist is about how easy you are to describe. Count the Yes answers across your fifteen Claude answers. Five or six means Claude can vouch for you. Three or four means it knows you exist but hesitates. Two or fewer means the web does not yet describe you well enough for a careful assistant to commit.

Check	Yes	Not yet usually means
Claude named you in at least half of your runs	You are describable and confirmed in more than one place	Claude cannot find you stated plainly on a source it trusts
It named you without hedging	Your details agree wherever Claude looked	Two sources disagree about your address, hours or services
Its sentence about you was specific and correct	Your website says what you do, for whom and where	Claude is paraphrasing a directory row rather than your own words
It cited a page that describes you, not only a listing	You have a quotable page	Your site is a brochure with no plain description Claude can restate
It named you for the specific-need question, not just the broad one	Your services are spelled out by name	Your services are implied, not stated
ChatGPT or Gemini also name you	Your clarity travels across the web	One clear page carries you and the rest of the web is silent

How Claude decides which businesses to name

When web search is available, Claude answers local questions by looking things up and drawing on what it can read there, then weighs what it finds before naming anyone. Without that access, it will say so rather than guess. What tends to earn a place is being confirmed in more than one spot: being described clearly and consistently across sources it can trust, rather than being loud in one.

The cautious streak matters here. Where a business is described vaguely, or the details conflict between one site and another, Claude is more likely to leave it out than to guess. It would rather name a business it is sure about than risk a recommendation it cannot stand behind.

What earns a mention in Claude

Because Claude rewards clarity and being easy to confirm, the work that improves your Claude visibility is the work that makes you easy to describe and hard to misread. These are the things that carry the most weight.

What Claude rewards	Why it matters	What to do
Clear description	Claude names what it can restate plainly	Say in plain words what you do, for whom, and where
Consistency	Conflicting details make Claude hold back	Keep your name, address and services identical everywhere
Being confirmed elsewhere	Being described across trusted sources builds confidence	Get listed and reviewed on the sites for your trade and town
Genuine reputation	A real, recent track record is easier to vouch for	Earn honest reviews rather than chasing a single mention

Why Claude may name fewer businesses than ChatGPT

If Claude names fewer businesses, or none, do not read it as Claude having less to say. Read it as a higher bar. A cautious assistant gives shorter lists, which means the businesses that do make it are the ones described clearly enough to be named with confidence. In one Peekl scan, an established clinic was named by Claude in 13 of 45 answers, more than Gemini's 6 and well ahead of ChatGPT's zero, even though its Google listings and reviews were solid. That gap is what a higher bar for naming looks like in practice.

The same audit shows why one check is not enough. Spread out, those 13 mentions covered 7 of the 15 questions, and just one question returned the clinic's name in every run. Ask once and you have a one-in-three chance of a misleading answer on most of the questions where Claude knows you at all.

For a borderline business, that caution is the difference between being included and being skipped. It is also good news: the same clarity and consistency that earn a place in Claude tend to help everywhere else too, so the work is not wasted on one assistant. For the wider method, see how to improve your AI search visibility, and for how the three assistants differ, see do ChatGPT, Gemini and Claude recommend the same businesses.

Bar chart from one Peekl scan: an established clinic was named by Claude in 13 of 45 AI answers, by Gemini in 6, and by ChatGPT in 0, showing Claude's higher bar to name a business. — One Peekl scan of an established London clinic: how often each assistant named it across 45 answers.

Checking ChatGPT, Gemini and Google AI Mode alongside Claude

Claude is the assistant the big industry studies skip. SOCi's 2026 index of 350,000 locations measured ChatGPT, Gemini and Perplexity, finding 1.2%, 11% and 7.4% of locations recommended, and did not include Claude at all. Our own audits are the only Claude data we have, which is one more reason to run your own check rather than assume.

Take the same five questions to ChatGPT, asking it which websites it used, and to Gemini, where the answer is mostly your Google Business Profile read aloud. Because Google AI Mode runs on the same signals as Gemini, what Gemini says about you is the nearest reading of AI Mode too. The free Peekl check asks ChatGPT or Gemini, your pick; Claude is in the full audit, which reads 45 answers from each of the three. For how rarely assistants name local businesses at all, see the local AI visibility guide.

Common questions

Does Claude recommend my business?

The only way to know is to ask it as a customer would, several times, with web search on. In our audits Claude names fewer businesses than ChatGPT or Gemini, so absence is common and not a verdict on your quality. Run the five-step check above and count how often your name appears across fifteen answers.

Why did Claude refuse to recommend anyone, or hedge its answer?

That is Claude's caution, not a dead end. It sets a high bar for naming a specific local business, and when it cannot be confident it may give fewer names or ask for more detail. It is a signal that the businesses in your area are not being described clearly enough for it to commit.

Does Claude browse the web to answer?

Usually, yes, when web search is switched on in your version of Claude. For local questions it will typically look things up and draw on what it can read there, then weigh it before naming anyone. So what the web says about you, and how clearly it says it, shapes whether Claude includes you.

Why does Claude ask where I am before it answers?

Because it will not guess a location for a recommendation it has to stand behind. Put your town in the question and it answers directly. When you check, do this every time, so each run is comparable and you are scoring the recommendation rather than the clarifying question.

How many times should I ask Claude before I trust the result?

Three runs per question across five questions is the minimum. In one Peekl audit Claude named a clinic for seven of fifteen questions at some point but for only one of them on every run, so a single answer, kind or unkind, is not a measurement.

Is being named by Claude harder than by ChatGPT?

Often, yes. Claude tends to give shorter, more cautious lists, so the bar to be included is higher. The upside is that the clarity and consistency that earn a mention in Claude usually help your visibility in other assistants as well.

Claude names my competitors but not me. What now?

It means your rivals are described more clearly and consistently across the sources Claude reads. The fix is to become just as easy to describe: sharpen what your website says, keep your details consistent, and earn genuine reviews, then check again over time.

See where you stand

Reading about AI search is useful; knowing your own place in it is what helps. A free Peekl check asks ChatGPT or Gemini, your pick, the questions your customers ask and shows whether it recommends you today, who it names instead, and how often. The full audit runs the same questions on all three assistants, Claude included.