Do Google reviews affect AI recommendations? Yes, here's by how much

Agencies have been telling SMB clients to collect reviews for years, mostly for the reputation benefit and a modest local ranking boost. Our latest research gives that advice a new, sharper edge: review volume is the single strongest signal correlating with whether ChatGPT and Perplexity surface a business at all.

The strongest signal in the entire dataset

In our AI Visibility Report, we analysed 10,000 US local businesses across ChatGPT and Perplexity, correlating AI outcomes against a wide range of SEO, local SEO, review and website signals. Out of all of them, one stood out clearly above the rest: Google Business Profile review count.

Businesses surfaced by both ChatGPT and Perplexity had an average of 133.4 reviews. Businesses not surfaced by either platform had an average of just 10.7. That's more than a 12x gap, and it's the strongest correlating factor we found anywhere in the data, the standout stat in our full AI visibility benchmark too.

The same pattern shows up again when we looked specifically at recommendation, not just awareness. Businesses that were "highly recommended" by AI (surfaced consistently, in a strong position, across five different queries) averaged 304.2 total reviews found across the web. Businesses that were never recommended averaged 117.5.

Review volume doesn't just correlate with whether AI knows about a business. It correlates with whether AI is willing to put that business forward as an answer.

It's not just volume, but the whole profile

Reviews were the strongest single factor, but they don't operate alone. A cluster of related Google Business Profile attributes moved in the same direction:

• Profile completeness: 77.8% for businesses found by both AI platforms, versus 31.5% for those found by neither
• Photos present: 88.9% versus 34.8%
• Claimed status: 91.5% versus 38.2%
• Average rating: 3.7 versus 1.7

Businesses surfaced by AI outperformed on every single one of these dimensions, not just the headline metric. That's a strong hint that AI systems aren't keying off one field. They're reading the whole profile as a composite signal of legitimacy and activity, and review volume happens to be the loudest part of that signal. That accumulation effect compounds fastest for bigger, more established businesses, the exact dynamic we cover in the scale advantage.

The twist: rating matters less than you'd think

Here's the part that surprised us. Star rating, the number most businesses (and most agencies) obsess over, is a comparatively weak predictor of whether a business gets recommended.

Across our recommendation buckets, average GBP rating barely moved: 3.8 for highly recommended businesses, 3.8 for frequently recommended, 3.8 for sometimes recommended, dropping only to 3.3 for businesses never recommended. Compare that to the near-3x spread in review count across the same buckets, and the pattern is clear: AI seems to care much more that a business has been reviewed a lot than that it's been reviewed perfectly.

That tracks with how AI sentiment behaves more broadly. Our reputation analysis found ChatGPT rates businesses positively almost 80% of the time regardless, and strongly negative responses were rare (0.5% of cases). A 4.8-star business and a 4.1-star business are both likely to get described warmly. What separates them in AI's eyes is closer to "how much evidence exists that this business is real, active and used by people", and volume is doing most of that work.

What this means for review strategy

If your review strategy has been built around chasing a marginally higher star rating, this data suggests a rebalance:

• Prioritise volume over polish. A business with 150 reviews at 4.2 stars is, on this data, more likely to be surfaced by AI than one with 20 reviews at 4.9 stars
• Treat review generation as an AI visibility lever, not just a reputation one. This is a genuinely new argument to bring to SMB clients who see reviews purely as a trust signal for humans
• Don't neglect the rest of the profile. Photos, completeness and claimed status all move with review count, so a review push paired with profile maintenance is stronger than either alone
• Don't over-index on rating. A near-perfect star rating with a thin review base is a weaker AI signal than a solidly-rated business with substantial review volume

Insites already tracks review volume and trend as part of every AI visibility report, so agencies don't need a separate reputation tool to make this case to a client.

Reviews were never just a reputation play. On this data, they're one of the clearest, most controllable levers a business has for whether AI recommends it at all.

This analysis is drawn from Insites' AI Visibility Report, based on a clean audit of 10,000 US local businesses across ChatGPT and Perplexity. Run a free AI visibility check on any business in seconds.