How Many Google Reviews Do You Need for AI Recommendations?

There is no minimum number of Google reviews for an AI recommendation, but there is a clear middle. Across 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, the median business the engine named had 226 Google reviews and the business it named first had a median of 342. Yet 23.5% of those answers still named at least one business with fewer than 20 reviews, while only 3.7% named any business under 4.0 stars.

That gives every branch two targets instead of one. I would hold the rating at 4.5 stars or better first, then close the gap to the review counts AI already shows in that trade and city, at a pace the team can keep without breaking Google's rules. The pace gets set at the end of a completed job, which is the moment review generation is built around.

Last verified: September 23, 2026.

No minimum, but a clear middle

Of the businesses Google AI Mode named with a rating, 8% had fewer than 20 Google reviews, 10.9% had 20 to 49, 27.9% had 50 to 199, 32.7% had 200 to 999, and 20.5% had 1,000 or more. The middle half ran from 74 reviews to 783. Counts above 999 are approximate, because Google AI Mode shows them rounded, as in 2.0K. Google itself says "More reviews and positive ratings can help your business's local ranking," and publishes no number, for Maps or for its AI.

So a new branch with a dozen reviews is not locked out by its count alone; nearly a quarter of answers included a business in that position. It is, however, competing in lists where the typical name carries a few hundred reviews, and that gap is what this article helps you size. Whether reviews matter to each engine at all is a separate question with a different answer per engine, covered in Do Google Reviews Affect AI Recommendations?

The rating floor is firmer than the review floor

Google AI Mode named thinly reviewed businesses far more readily than low-rated ones. Among the 8,523 answers that showed at least one rated card, 23.5% named a business with fewer than 20 reviews and 14.6% named one under 4.5 stars, but only 3.7% named one under 4.0. Across every rated business it named, 83% carried 4.8 stars or more, 3.5% sat between 4.0 and 4.4, and 1.0% were under 4.0. The median was 4.9.

We don't see the businesses it passed over, so this doesn't prove that Google AI Mode screens out low ratings. What it does show is the company a 4.2-rated branch keeps on these lists: more than four in five of the businesses named beside it are at 4.8 or higher.

The weekly checks point the same way. Among businesses with 100 or more reviews, those rated 4.0 to 4.4 were named in all three weeks 6.7% of the time, against 14.8% at 4.8 and 15.8% at 4.9.

That sets the order of work. A branch at 4.3 with 400 reviews has a service problem before it has a volume problem, because asking more customers at the same service level mostly adds more reviews at the same average. Read its last 20 reviews, find the complaint that keeps coming back, fix it in the field, and reply to each of those reviewers before you push the count.

A perfect 5.0 is not the goal

Businesses showing a perfect 5.0 were not named more consistently than 4.9-rated ones. Among businesses with 100 or more reviews, 14.5% of those at 5.0 and 15.8% of those at 4.9 were named in all three weeks. The raw numbers make 5.0 look worse (11.4% against 14.9%), but most of that gap is volume: the median 5.0-rated business had 101 reviews and the median 4.9-rated business had 418.

Nothing here says a 5.0 hurts. It says a 5.0 on a small count doesn't stand in for a larger count, and chasing the last tenth of a star is a poorer use of a branch manager's week than asking every customer. For context, 28% of the rated businesses Google AI Mode named showed exactly 5.0 and 35% showed exactly 4.9.

The bar moves with the trade and the country

The middle changes a lot by trade. For the nine HVAC questions we ask each week, the businesses Google AI Mode named had a median of 449 Google reviews. For the nine plumbing questions it was 364, for the two roofing questions 130, and for the six electrical questions 113. Electrical answers were also the most open to small businesses: 40.8% named at least one business with fewer than 20 reviews, against 16.9% for plumbing and 11.6% for HVAC.

The other trades rest on one question each, so read them as answers to that question and nothing wider. For "who are the best pest control companies for home pest treatment?" the median named business had 412 reviews. The single questions on garage door repair, duct cleaning, and window and door replacement came in at 280, 209 and 169; water damage restoration at 80; lawn care and landscaping at 67.

Canada runs lower. The median business Google AI Mode named in our 40 Canadian metros had 134 reviews, against 332 in our 60 US metros, and 11.3% had fewer than 20 reviews, against 5.8% in the US. A Canadian branch that benchmarks itself against a US number sets a target it doesn't need.

More reviews, more staying power

AI answers change from week to week. Ask Google AI Mode the same question for the same city a week later and about 4 in 10 of the businesses it named come back, so the more useful question is whether a business keeps getting named. Among businesses Google AI Mode named at least once for a question and city, those with 1,000 or more reviews came back in at least two of the three weeks 43.3% of the time. Businesses with fewer than 20 reviews came back 25.7% of the time, and each band in between fell in order: 30.4% for 20 to 49, 36.5% for 50 to 199, and 41.1% for 200 to 999.

Order in the answer moved the same way. In answers that named three or more rated businesses, the first business named had a median of 342 reviews, the second 237 and the third 195, and the most-reviewed business came first 36% of the time against 25% expected by chance. Review count also tracks how big and how old a company is, so treat this as a pattern among named businesses, not proof that more reviews move a branch up the list.

What ChatGPT and Gemini show

Gemini's business cards told a similar story. The median business it named with a card had 254 reviews, 73% were rated 4.8 or higher, and 2.1% were under 4.0. We only captured Gemini's cards from September 14, and not every Gemini answer shows them, so treat Gemini as a second opinion rather than a benchmark. Where reviews sit among the other factors we measured, engine by engine, is in AI local ranking factors for 2026.

ChatGPT's map needs a warning label. About one in three of its rated business cards linked to a Yelp page, and those cards carried far fewer reviews: a median of 32, against 168 for its other cards. For businesses we could match across engines, the other cards sat within 0.1 star of Google AI Mode's rating 98% of the time. If a branch looks thin or low-rated on ChatGPT, check whether its card links to Yelp before you draw any conclusion about its Google reviews; why ChatGPT shows a lower star rating than Google walks through it.

Set a review pace for each location

This is how I would turn the benchmark into a number a regional manager can own. Start with the trade median above, or better, the counts on the cards Google AI Mode shows for your top question in that city. Subtract the branch's current count, then divide by the number of months you're willing to wait.

For a sense of what's realistic, the median company in our customer panel collected 9.7 new reviews per location a month over the 90 days ending September 2, 2026, and the 75th percentile company collected 26.4. That panel is 11 home-services companies, aggregate only, published in the Home Services AI Visibility Index.

The arithmetic is humbling in some trades and easy in others. An HVAC branch with 150 reviews is 299 short of the 449 median; at 9.7 a month that takes about 31 months, and at 26.4 about 11. An electrical branch with 40 reviews is 73 short of 113, which is about seven and a half months at the median company's pace. Same brand, very different asks. When the pace a branch needs is above 26.4 a month, stretch the window to two years rather than pushing the team toward shortcuts.

Set the target on the location's process, not on individual technicians. Google's review policy prohibits "Merchants requesting that staff solicit a certain number of reviews," along with incentives of any kind and selectively asking happy customers. What a manager can own is the share of completed jobs that get a neutral request, measured by branch, and Google's own help page says a business can "ask customers to visit a Google link or scan a QR code." The compliant review collection playbook has the full policy line.

What to do this week

Pick your five weakest branches by rating, not by count, and give each one a short worksheet:

• Record the Google rating, the review count, and the date the most recent review arrived.
• If the rating is under 4.5, spend the week on the complaint that repeats in the last 20 reviews, and reply to each of those reviewers.
• If the rating is 4.5 or better, ask Google AI Mode the question customers ask most for that branch's trade and city, and write down the review counts on the cards it shows.
• Set a 12-month count target from those cards, divide the gap by 12, and compare the monthly pace with 9.7 and 26.4.
• Put a neutral ask on every completed job at that branch, and track the share of jobs that got one.

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.

The review counts and ratings in this article are the figures Google AI Mode displayed on its business cards for businesses also named in the answer text: 34,017 rated business appearances across 8,523 answers with a rated card. Google AI Mode's cards can hide business names in about 5% of answers. Weekly persistence covers every question-and-city pair that returned an answer in all three weeks, with each business placed in the band of the highest review count Google AI Mode showed for it in that metro. Gemini figures cover place cards from September 14 to 22, 2026, and ChatGPT figures cover map cards from September 18 to 22, 2026. The customer-panel review pace is a separate Cheers production pull, described on the methodology page.

Sources

• Cheers Market Baseline, August 31 to September 20, 2026, 34,742 answers from four engines in 100 US and Canadian metros. Cheers Research methodology. Source of every Google AI Mode, Gemini and ChatGPT figure above.
• Home Services AI Visibility Index 2026. Cheers Research, edition September 2, 2026, corrected September 23, 2026. Source of the 9.7 and 26.4 new reviews per location a month across 11 home-services companies.
• Prohibited and restricted content. Google Maps user-generated content policy, undated, read September 23, 2026. Source of the prohibitions on staff review quotas, incentives, and selectively soliciting positive reviews.
• Tips to get more reviews. Google Business Profile Help, undated, read September 23, 2026. Source of the guidance that a business can ask customers to use a Google link or QR code.
• Tips to improve your local ranking on Google. Google Business Profile Help, undated, read September 23, 2026. States that more reviews and positive ratings can help a business's local ranking.

Dylan Allen-Arnegård is the CEO & Co-Founder of Cheers, the local search platform for multi-location service businesses.