How to check if your business shows up when customers ask AI

Why one screenshot will lie to you

The single most common mistake is asking an engine once and treating the answer as the truth. AI engines are nondeterministic: ask the exact same question twice, a minute apart, and you routinely get a different set of businesses named. In our own run-to-run data, most brand names that appear in one answer are gone by the next ask (the numbers are here). So a single screenshot showing your name is false comfort, and one showing only competitors is false panic. The only honest measurement is a frequency — across many runs, on each engine, how often are you named, and who gets named instead of you.

That's the whole trick to checking yourself accurately: don't ask once. Ask the same questions several times, on more than one engine, and count.

The free DIY method, step by step

You need the four engines your customers actually use — ChatGPT, Perplexity, Gemini, and Claude — each open in its own tab. Free accounts are fine. Then:

1. 1Write down the questions a real buyer would ask — not "is [your business] good" (that leads the engine to you). Ask the way a stranger with a wallet would: "best {your category} in {your city}", "who are the top {category} near {neighborhood}", "I need a {category} in {city} — who should I call?". Aim for 3–5 buyer-intent questions.
2. 2Start every question in a fresh chat. Open a new conversation each time so the engine isn't primed by what you asked before. A follow-up in the same thread is contaminated — it remembers you.
3. 3Ask each question 3 times, on each engine. Yes, it's repetitive — that repetition is the measurement. Three runs × four engines × (say) four questions is 48 answers. That's enough to see a pattern instead of a coin flip.
4. 4For every answer, record two things: was your business named (yes/no), and which competitors were named. A tally sheet or a spreadsheet with one row per run is plenty.
5. 5Turn off personalization where you can. Use a logged-out or incognito window when the engine allows it, so you're seeing what a stranger sees, not an answer shaped by your own history and location.

How to turn your tallies into a score you can trust

Once you've run the grid, three numbers tell you almost everything:

• Your mention rate — of all the answers you collected, what fraction named you? Named in 6 of 48 answers is a 12.5% mention rate. This is your single most important number, and it's the one no screenshot can give you.
• Your per-engine split — mention rate per engine, not just overall. Engines disagree hard; you might be at 40% on Perplexity and 0% on ChatGPT. Averaging them hides the engine that's quietly costing you customers.
• Your top competitor — the business named most often instead of you. Count how many answers named them versus you. If a rival appears in 20 answers and you appear in 6, that's the number that should keep you up at night: same market, same questions, three times your visibility.

There's no official pass mark, but our benchmark data gives you a yardstick. When we ran this same measurement rigorously on 10 established US local businesses, the average mention rate was 38% and the average visibility score was 20 out of 100 — and these were firms with polished sites and page-one Google rankings. If your DIY mention rate is well under a third, you're not an outlier; you're normal, and that's exactly the problem worth fixing.

What the numbers usually reveal

Three findings show up again and again once people actually measure instead of guessing — and they're the reason a real check is worth the fifteen minutes:

• Google rank doesn't carry over. In our 10-business benchmark, several firms ranking on page one of Google still scored 13–28 out of 100 in AI answers. Engines synthesize from directories, "best {category} in {city}" roundups, and cross-platform review consensus — a different surface than the one your SEO optimizes. Being invisible in AI while winning on Google is the single most common surprise.
• A competitor is usually named more than you. In 6 of our 10 audits, one rival was named more often than the business itself; in 9 of 10, at least one competitor tied or beat it. AI answers aren't empty — they recommend someone, specifically and by name, and by default that someone isn't you.
• Markets are winner-take-most. Engines tend to converge on one or two "consensus picks" per local market and mention everyone else thinly. If the consensus pick in your city isn't you, every AI-assisted buyer is being steered to the same competitor, every day.

Methodology (so you can trust the yardstick above)

The benchmark figures cited here come from AskedAbout's own study of 10 US local businesses (5 personal-injury law firms, 5 med spas) across 9 metro areas, audited 2026-06-12 via our production pipeline: buyer-intent questions per business submitted to ChatGPT, Perplexity, Gemini, and Claude; each answer parsed for business and competitor mentions by a judge model; the 0–100 score combines mention rate, prominence, and consistency across runs. Audited businesses are anonymized — this is benchmark data, not a callout. n=10, and we say so; we re-publish as the sample grows. Full write-up: the local-business AI visibility benchmark.

DIY vs. an automated check — an honest comparison

The manual method is genuinely useful — do it, especially if you've never looked. Its limit is patience and consistency: 48 hand-tallied answers is a real signal, but 300 answers scored identically every time is a baseline you can re-measure against to see whether your fixes worked. That's the whole reason AskedAbout exists. The free 60-second check gives you your score, mention rate, and top competitor without the spreadsheet; the $79 one-time Full Audit adds the per-engine share-of-voice, the sources the engines are citing about your market, and a ranked fix plan. One payment, no subscription — and we query the official APIs and say so, no scraped single screenshots.

See also

• Once you know your number, fix it: how to show up when ChatGPT recommends a business like yours and why AI keeps recommending your competitors.
• The data behind the method: why AI answers change every time you ask, the 10-business benchmark, and which engine recommends businesses least.
• Your industry: AI visibility for med spas · law firms · dentists · SaaS · home services.

FAQ