28 of 63 local businesses showed up in all three places customers look.
This is a monthly series, not a one-off study. Every month we draw a fresh sample of 63 real local businesses, none of them our clients, and measure the three surfaces that decide whether they get found: Google search, the map pack and the AI answers. The method is identical every month, so the editions are comparable readings of the same population. They are not the same businesses tracked over time, and at this sample size a single month’s figure carries roughly ±12 points of sampling error. We publish that alongside the number.
28 of 63 appeared on all three surfaces.
Of the 63 local service businesses sampled in August 2026, 28 were findable in Google’s top 20 organic results, the local map pack and at least one AI assistant for their core buying query. 18 appeared in exactly one of the three. No business in this month’s sample was invisible on all three. On a sample this size that 44% carries a margin of error of about ±12 points.
Here is the full distribution. Every business was measured on all three surfaces, so nothing below is estimated or extrapolated.
Last month’s edition put the all-three figure at 47%. The two samples share no businesses at all: each edition draws fresh, so this is two independent estimates of the same population rather than the same companies tracked over time. That is a 3-point difference against a 17-point margin of error on the comparison, so the two readings are statistically indistinguishable. Nothing detectable changed. We are reporting that plainly rather than dressing it as movement.
How the Index works.
63 businesses. Eight trades this month: landscaping, roofing, HVAC, tree service, paving, window tinting, cabinets and water restoration. Four metros: Lexington KY, Murfreesboro TN, Richmond VA and Spokane WA. Two businesses per trade per metro, drawn from live Google Maps results for each trade’s core buying query.
The sample rotates. Trades and metros step through fixed pools of twelve each month, so no two consecutive months share the same mix and the Index never settles into one region’s quirks. We exclude our own clients from every draw, so nothing here is flattered by work we did.
That draw method has a consequence we state rather than bury. Because the sample comes from map results, near-universal map presence is a property of the method, not a finding. It is also what makes the rest meaningful: these are businesses that already cleared the bar most local SEO advice is about.
For each business we measure three things. Whether the domain ranks in the top 20 organic results for its core service and city. Whether it appears in the local map pack for that query. And whether ChatGPT, Claude, Gemini or Perplexity name it in response to a buyer-intent prompt.
An engine that errored is never counted as a business not being found. Counting API failures as absences would manufacture a more dramatic number than the true one, so errored cells are tracked separately and excluded from the denominator.
1 of the 64 businesses we drew are excluded from the figures above, because their names are their category. A company literally called “Tucson Plumbing” cannot be told apart from any other Tucson plumber by reading an AI answer: the words appear whoever the answer actually recommends. Scoring them as found would inflate our own headline, and scoring them as missing would deflate it, so they are reported here and left out of the rates. That leaves 63 businesses measured on all three surfaces.
Each surface on its own.
| Surface | Found | Rate |
|---|---|---|
| Google organic top 20 | 37 of 63 | 59% |
| Local map pack (top 3) | 62 of 63 | 98% |
| Named by at least one AI tool | 37 of 63 | 59% |
Each engine named a similar share of the sample. Same prompt, four engines:
The four AI tools mostly do not name the same businesses.
Those rates look similar, and that similarity is misleading. The engines land on close to the same NUMBER of businesses while disagreeing about WHICH ones, and that is only visible if you check the overlap.
37 of the 63 businesses were named by at least one engine. Of those, 12 were named by exactly one of the four, and only 7 were named by all four. Agreement between any two engines ran from 37% to 58%.
The practical consequence is the part worth keeping. “Am I visible in AI” is not one question, it is four, and the answer you get depends heavily on which tool you happened to open. Checking the one you have heard of will mislead you in either direction, and a business that appears in ChatGPT has no particular reason to appear in Perplexity.
Window Tinting led this month. Tree Service trailed.
| Trade | On all three | Rate |
|---|---|---|
| Window Tinting | 7 of 8 | 88% |
| HVAC | 6 of 8 | 75% |
| Cabinets | 4 of 8 | 50% |
| Water Restoration | 3 of 7 | 43% |
| Landscaping | 3 of 8 | 38% |
| Paving | 3 of 8 | 38% |
| Roofing | 1 of 8 | 12% |
| Tree Service | 1 of 8 | 12% |
Our own score this month: 12 out of 100.
Every month the Index includes Tristar measured with the exact same probes, on the buyer questions we would want to be found for. This month 2 of the four AI tools named us and 2 cited our website, which works out to 12/100 on the same scale we use for everyone else.
When we first measured ourselves in June 2026, on a similar set of buyer questions, the score was zero. We publish this row every month because we sell this work, and a scoreboard that skips the scorekeeper is not one you should trust. If our number stalls, you will see that too.
What this month’s reading does not show.
Stated up front because a study that only lists its strengths is marketing wearing a lab coat.
63 is a small sample. Enough to establish a pattern, not enough for precision. Per-trade cells are directional only.
The sample rotates monthly. That keeps the Index honest across markets, and it means month-to-month movement mixes real change with sample change. Read the direction over quarters, not the wiggle between months.
Map presence is a selection artifact. Businesses were drawn from map results, so their map rate describes our method, not the market.
AI answers vary between runs. These systems are non-deterministic. We measure one run per engine per business, every month, the same way.
Name matching has a floor. We detect a business by its name and domain in an answer. Businesses named after their own category cannot be detected this way at all, so we exclude them and say how many rather than guessing in either direction.
Visibility is not revenue. We measure whether businesses can be found, not whether they are busy.
Check yourself before you believe anybody.
The useful thing here is the method, not our number. Search your core service and your town in an incognito window and note whether you appear in the map pack and the organic results. Then ask ChatGPT, Perplexity, Gemini and Google’s AI results who they would recommend for that service in your town, phrased the way a customer would.
Then count your surfaces. Three and you are in the 44%. One, and you now know something about your business that most of your competitors do not know about theirs.
Our free check automates exactly this across all four AI tools plus Google and the map pack, and texts you a link to the result. The manual version costs nothing and we would rather you know either way.
Citing the August 2026 Index.
You are welcome to cite any of this, including competitors. We ask only that you link the Index rather than the summary, and quote the sample size next to the figure.
Copy this citation:
Tristar Marketing Solutions, “Tristar Local Visibility Index, August 2026.” 28 of 63 sampled local businesses appeared in Google’s top 20, the local map pack and at least one AI assistant. https://old.tristarmarketingsolutions.net/research/visibility-index-2026-08/
The anonymized measurements are available on request: every business is reported by trade and metro only, and nobody in the sample is named.
If you want the underlying numbers for a specific finding, or you think we have made an error, ask. We would rather somebody check our arithmetic than take it on faith.
Run this month’s Index on your own business.
Free, 30 seconds, the same three surfaces. See who gets the call instead of you, and a link to your result you can keep. No sales call.
Check My Business Free →