Original research
Google’s top results and AI’s recommendations are not the same list
We measured 38 local service markets across 50 scans and 350+ businesses. In none of them did Google’s local results and the businesses AI recommends match. Four out of five businesses AI names do not appear in Google’s local pack for the same search at all.
Most local businesses still treat “show up on Google” and “get recommended by AI” as the same job. If that were true, the two lists would broadly agree — the businesses Google ranks first would be the ones ChatGPT and Perplexity name when a customer asks for a recommendation.
We had the data to check, so we did.
What we found
Four out of five businesses that AI recommends are not in Google’s local pack for the same query. Across every business AI named in these scans, roughly 80% did not also appear in Google’s top local results for the identical search.
Not one market agreed. We counted a market as “agreeing” only if the two lists matched name-for-name in both directions — no business on either list missing from the other. Of the 38 markets measured, that happened zero times. Not once.
The practical version: the business winning Google in your town is usually not the business ChatGPT recommends there. These are two separate races, and most owners are only running one of them.
Why this happens
Google’s local pack is driven heavily by proximity, Google Business Profile completeness, and review volume — signals tied to a map. An AI assistant answering “who’s the best plumber in Denton” is doing something different: it is summarising what the web says about businesses. It pulls from directory profiles, roundup articles, forum threads, and the businesses’ own pages, then names whoever the sources it trusts describe most clearly and most favourably.
Those are different inputs, so they produce different lists. A business with a strong Google Business Profile and thin written presence wins the map and loses the answer. A business that is written about clearly in the places models read can be named by AI while sitting on page three of Google.
How we measured it
Every data point here comes from real scans we ran through our own product on real businesses — no survey, no panel, no third-party dataset.
For each market we issued a natural buying query in the form “best {trade} in {city}”, then captured two lists for that same query:
- The Google list — the businesses appearing in Google’s local results.
- The AI list — the businesses named in answers from live AI assistants.
We then compared them name-for-name. A business counted as “AI-only” when it appeared in the AI answer but nowhere in Google’s list for the same query. Scans where either list came back empty were excluded entirely, so an engine outage can never be read as “disagreement.”
| Measure | Value |
|---|---|
| Scans with both lists populated | 50 |
| Distinct market queries | 38 |
| Distinct businesses observed | 350+ |
| AI-named businesses absent from Google’s list | ~80% |
| Markets where both lists matched | 0 |
The percentage is rounded to the nearest 5% and the business count is floored to the nearest 25, deliberately. These are honest estimates from a real sample, not a claim of laboratory precision.
What this study does not show
- The sample is geographically concentrated. Most markets are in North Texas — Denton, Frisco, Flower Mound, Lewisville, Carrollton, Little Elm and neighbouring cities — with a smaller number in Dallas, Houston, Denver, Phoenix, Indianapolis and Kansas City. We have not shown this holds nationally.
- It covers local service businesses only — roughly eleven trades including insurance, HVAC, plumbing, roofing, pest control, landscaping, garage doors and foundation repair. It says nothing about e-commerce, B2B software, or national brands.
- Each market was measured at a point in time. AI answers vary between runs and change as models update. We have not tested how stable an individual answer is week to week.
- It shows divergence, not cause. We can show the two lists differ. We are not claiming to have proven why any specific business appears on one and not the other.
- It is our own methodology measuring our own scans. We think that is the honest way to state it rather than dressing it up as an independent study.
What a business owner should take from this
Not “abandon Google.” Google’s local pack still drives a large share of local demand and will for a long time.
The useful conclusion is narrower and more actionable: your Google ranking tells you almost nothing about whether AI recommends you. If you have never checked the second list, you genuinely do not know whether you are on it — and in our sample, the odds were four to one that a business AI named was invisible in Google’s results for that same search, and by extension that plenty of businesses ranking well on Google were absent from the AI answer.
Checking takes a couple of minutes. Ask ChatGPT or Perplexity for the best business in your category and city, and read the names it gives back. If yours is not among them, that is not a Google problem and more Google work will not fix it.
See both lists for your market
Free scan, no card. We run the buying questions customers actually ask in your city, then show you who AI names, who Google ranks, and where the two disagree.
Run my free scan →