To show up in Google AI search, a local business has to look like the businesses Google AI Mode already names: rated 4.8 stars or higher, backed by a real review history, and linked from a Business Profile to a working website and booking page. In 9,435 Google AI Mode answers to 32 home-services questions in 100 US and Canadian metros, collected by Cheers between August 31 and September 20, 2026, only 3.7% of the answers that showed rated business cards named any business under 4.0 stars, while 23.5% named at least one business with fewer than 20 reviews. For a multi-location brand, that pattern puts a branch sitting at 4.2 stars further from the businesses Google AI Mode names than a new branch with 15 good reviews.
Last verified: September 23, 2026.
Important
Every number here describes the businesses Google AI Mode named, not a rule Google applies. Google publishes no local recommendation formula, so read each finding as an association in our sample.
In this article, Google AI search means Google AI Mode, the conversational search mode, and every number comes from it. AI Overviews, the summaries at the top of regular results, are a separate feature that Google says "may use different models and techniques", so we keep our data claims to AI Mode. If you want a one-location starting point on the other engines, the free AI Visibility Grader checks a business across ChatGPT, Gemini and Perplexity.
What Google says, and what it leaves out
Google's guide to its AI features in Search is short on local detail and firm on one point:
Our generative AI features on Google Search are rooted in our core Search ranking and quality systems.
The same guide, last updated July 10, 2026, says a page must be indexed and eligible to show a snippet to appear in those features, that "you don't need to create new machine readable files, AI text files, markup, or Markdown", and that "using products like Merchant Center and Google Business Profiles can help your products and services be visible in both AI responses and other Google Search results."
For a local business, that boils down to normal Search eligibility plus an accurate Business Profile. The guide sets no standard for ratings, review counts or booking links, and the answers themselves say a lot about all three.
What Google AI Mode put in front of customers
A Google AI Mode answer to a home-services question is, nine times in ten, a short list of business cards. It showed rated business cards in 90% of the answers we collected, and the share barely moved with the kind of question: 89.7% for "best company for this job" questions, 91.6% for emergencies, 90.0% for same-day requests, 90.9% for questions about reviews and fair pricing, and 93.3% for deals and quotes. A typical answer carried about four cards, and Google AI Mode showed a star rating and review count for 90% of the businesses it named.
Who Google AI Mode named
4.9 stars, 226 reviews
4.9
Median star rating of the businesses Google AI Mode named with a rating
Base: 34,017 rated business appearances in 8,523 answers
83%
Rated 4.8 stars or higher
34,017 rated appearances
1.0%
Rated under 4.0 stars
34,017 rated appearances
226
Median Google review count
34,017 rated appearances
Cheers Market Baseline, August 31 to September 20, 2026: 34,742 answers from ChatGPT, Gemini, Perplexity and Google AI Mode to 32 home-services questions in 100 US and Canadian metros. Aggregate only.
| Finding | Value | Base |
|---|---|---|
| Median star rating of the businesses Google AI Mode named with a rating | 4.9 | 34,017 rated business appearances in 8,523 answers |
| Rated 4.8 stars or higher | 83% | 34,017 rated appearances |
| Rated under 4.0 stars | 1.0% | 34,017 rated appearances |
| Median Google review count | 226 | 34,017 rated appearances |
The businesses on those cards were strongly rated. Across 34,017 rated business appearances in 8,523 answers, the median rating was 4.9 stars, 83% were rated 4.8 or higher, and 1.0% were under 4.0. Review counts spread much wider: the median business showed 226 Google reviews, 8% had fewer than 20 and 21% had 1,000 or more.
The rating floor is firmer than the review floor
The chart at the top of this page puts those two distributions side by side at the answer level. Among the 8,523 answers that showed rated cards, 23.5% named at least one business with fewer than 20 reviews, 14.6% named one rated under 4.5 stars, and just 3.7% named one under 4.0.
A thin review count did not keep a business off the list nearly as often as a weak rating did. That is useful for a new or acquired location: a branch with a short, strong review history was still being named. It is a warning for an established branch that has slid into the low 4s, because Google AI Mode rarely named businesses there. We can't tell whether Google screens them out or whether those businesses simply lose on other grounds, since we only see who was named.
The review floor also moves by trade and country. In electrical answers, 40.8% named at least one business with fewer than 20 reviews, against 11.6% of HVAC answers, where the median named business showed 449 reviews. Canadian businesses showed fewer reviews than US ones: a median of 134 against 332.
The first business named usually has the most reviews
Order carried a signal the rating didn't. In Google AI Mode answers that named three or more rated businesses, the first business named had a median of 342 reviews, against 237 for the second and 195 for the third. The most-reviewed business came first 36% of the time, where chance would put it first 25% of the time. Average ratings barely moved from first place to third.
Google AI Mode, answers naming three or more rated businesses
The first name had the most reviews
36%
Most-reviewed named first
Median Google reviews by position in the answer
Cheers Market Baseline, August 31 to September 20, 2026: 34,742 answers from ChatGPT, Gemini, Perplexity and Google AI Mode to 32 home-services questions in 100 US and Canadian metros. Aggregate only.
| Position in the answer | Median Google reviews |
|---|---|
| Named first | 342 |
| Named second | 237 |
| Named third | 195 |
That is a correlation across 7,873 answers, and review count also tracks how big and how established a business is. It still gives an operator a clear read: once the rating clears the high 4s, review depth is the clearest difference between the first name and the third.
Almost a third of answers linked to a booking page
The booking-link finding is the easiest to act on. In all, 31.2% of Google AI Mode answers (2,947 of 9,435) linked directly to a business's online booking or scheduling page hosted on field-service software. Housecall Pro pages appeared in 20.5% of answers, ServiceTitan in 10.6% and Jobber in 3.1%. ChatGPT and Gemini never linked one, and Perplexity did in 0.2% of answers.
Google AI Mode links to booking pages
Almost a third linked a booking page
0
ChatGPT or Gemini answers that did
Answers linking an online booking page
2,947 of 9,435 answers
Cheers Market Baseline, August 31 to September 20, 2026: 34,742 answers from ChatGPT, Gemini, Perplexity and Google AI Mode to 32 home-services questions in 100 US and Canadian metros. Aggregate only.
| Questions | Share of answers with a booking link |
|---|---|
| All 32 home-services questions | 31.2% (2,947 of 9,435 answers) |
| Plumbing, 9 questions | 39.4% |
| HVAC, 9 questions | 38.9% |
| Electrical, 6 questions | 33.3% |
| Roofing, 2 questions | 1.8% |
The share depends on the trade: 39.4% of plumbing answers, 38.9% of HVAC answers and 33.3% of electrical answers carried a booking link, against 1.8% of roofing answers, which fits a trade where the first step is usually an inspection or estimate rather than a booked service call. US answers carried one more often than Canadian answers (33.4% against 27.9%).
Where do those links come from? Our best guess is the booking link on each business's Google Business Profile, but that is an inference we have not verified. What Google does document is how that link works. A Business Profile can carry up to 10 links for booking an appointment, and Google's business links policy sets the conditions:
Local business links must lead to a dedicated landing page for your business. For businesses with multiple locations, action links must lead to a website for a specific location.
The same policy says the business must be verified before it can add a link, that the link must let a customer complete the action it promises, and that Google "may remove any non-compliant links." For a franchise or roll-up, that rules out pointing every location at one corporate scheduler page. Each profile needs its own branch's booking page, and it needs to work.
Emergency answers lead with 24/7
For emergency questions ("right now", "stopped working", "available today"), 92.8% of Google AI Mode answers mentioned 24/7 or after-hours availability, against 48.0% for "best company for this job" questions. That count includes any mention, generic advice included, so it shows what the answer talks about rather than why a business was picked.
It still tells you where a wrong hours listing hurts most. If a location takes calls at night, its Business Profile hours, its website and the person answering the phone should all say so. If it doesn't, don't claim it. The emergency hours guide covers where to publish that promise.
Next step
Is AI recommending your business?
Find out how visible you are across ChatGPT, Gemini, Perplexity, and AI Overviews.
Google AI Mode reads Google and your website
Google AI Mode's citations looked nothing like ChatGPT's. It cited Google's own properties in 95.5% of its answers and businesses' own websites in 92.5%, and those two groups took about 43% and 34% of all its citations. Only 38.3% of its answers cited any directory, review platform or "best of" list, against 84.2% for ChatGPT.
When Google AI Mode linked to a contractor's website, 66.9% of those links went to the homepage and 13.5% to a location or city page. We think the homepage links come from the website button on the Business Profile, again an inference. If that is right, a multi-location brand whose profiles all point at the corporate homepage is sending Google AI Mode to a page that can't describe any one branch. Linking each Business Profile to its location page fixes that. Service pages drew only 5.0% of Google AI Mode's business-website citations, against 26.5% on Perplexity, so build them for the job and the other engines, and don't expect them to carry Google AI Mode alone.
The profile carries more weight on Google AI Mode than anywhere else we measured. Whether it is enough for AI visibility on its own is a separate question, and the answer is no once ChatGPT is in the picture: ChatGPT cited Google's own properties in 0.1% of its answers.
Ask for reviews the way Google allows
A high rating on a real review history comes from asking every customer, not from filtering who gets asked. Google's contribution policy prohibits businesses that "selectively solicit positive reviews from customers", bans incentives for reviews, and lists "merchants requesting that staff solicit a certain number of reviews" as prohibited. Ask every customer the same neutral way, and let the rating be what it is.
What to do this week
Pull every location's Google star rating and review count into one sheet, then work down this list:
- Flag every location under 4.5 stars, read its last 20 reviews, and fix the operational problem they describe, because more reviews won't lift a rating the service keeps earning.
- Check the booking link on every Business Profile. It should open that location's own booking page on your field-service software, and a test booking should go through.
- Compare each profile's hours with what the phone line actually does after 6 p.m., and remove any 24/7 claim the branch can't answer.
- Point each profile's website link at the location page, and make sure that page names the services and areas the branch covers.
- Train every technician and front-desk person to ask every customer for a Google review the same way.
For the ChatGPT side of the same job, the sources, the Yelp cards and the directories, read how to get ChatGPT to recommend your business. The businesses in these answers also change from week to week, which is why AI recommends different businesses every time you check. The 2026 AI local ranking factors put all four engines on one page.
Methodology
Every week, Cheers asks ChatGPT, Gemini, Perplexity and Google AI Mode the same 32 home-services questions in 100 metro areas across the United States and Canada, and records every business each engine names, with its star rating, review count and map position where the engine shows one, and every source it cites. This analysis covers the three weeks from August 31 to September 20, 2026: 34,742 usable answers (9,434 from ChatGPT, 9,435 from Google AI Mode, 9,364 from Gemini and 6,509 from Perplexity) across 60 US and 40 Canadian metros. We excluded empty or cut-off answers, most of them from a Perplexity collection outage. Each question was asked once per city per week, so week-to-week changes include normal answer-to-answer variation. Business names were extracted automatically; a hand check of 40 answers found the extraction caught about three in four of the businesses a reader would count, so counts are lower bounds. Star ratings and review counts are the figures each engine displayed; for ChatGPT and Gemini we captured them from September 18 and September 14 onward. Findings describe associations in this sample, not how any engine ranks businesses.
For this article: ratings and review counts are what Google AI Mode displayed on its place cards, and it shows counts above 999 rounded (for example 2.0K). Its place cards hid business names in about 5% of answers. Booking links are citations to booking pages hosted on field-service platforms (Housecall Pro, ServiceTitan, Jobber and others). Directory, platform and page-type groups come from hand-built domain lists and URL rules; "business website" counts only domains we could match to a named business, so it is a lower bound. Trade figures quoted here come from trades with two or more questions.
Sources
Checked September 23, 2026.
- Optimizing your website for generative AI features on Google Search. Google Search Central, last updated July 10, 2026. Source of the core-systems quote, the snippet eligibility rule, the quoted line on machine-readable files and the quoted line on Business Profiles.
- AI features and your website. Google Search Central, last updated December 10, 2025. Source of the quoted line that AI Mode and AI Overviews may use different models and techniques.
- Manage your local business links. Google Business Profile Help, no date shown, checked September 23, 2026. Source of the up-to-10 booking links per category.
- Business links policies and guidelines. Google Business Profile Help, no date shown, checked September 23, 2026. Source of the quoted location-specific landing page rule, the verification requirement, the complete-the-action rule and link removal.
- Set up bookings through a provider. Google Business Profile Help, checked September 23, 2026. Background on Reserve with Google providers and adding your own booking link.
- Prohibited and restricted content. Google Maps User Generated Content Policy Help, checked September 23, 2026. Source of the quoted selective-solicitation and staff-target rules and the incentive ban.
- Cheers Market Baseline, August 31 to September 20, 2026: 9,435 Google AI Mode answers within 34,742 answers from four engines to 32 home-services questions in 100 US and Canadian metros, aggregate only. Research methodology. Source of every card, rating, review, order, booking, 24/7, citation and page-type figure.
Dylan Allen-Arnegård is the CEO & Co-Founder of Cheers, the local search platform for multi-location service businesses.