Why five-star reviews with no text are nearly worthless now
For fifteen years the advice was simple: get more reviews, keep your rating high. Volume and stars. That was the game, and it worked because a human scanning a map pack sees a number and a count, makes a snap judgement, and moves on.
An AI assistant does not do that. It reads them.
What changes when the reader is a machine
Consider two reviews, both five stars.
"Great service, highly recommend!"
"Brought my BMW in for a coolant leak on Thursday, they quoted $340 and charged $340, had it back Friday morning. Explained what was wrong without being condescending about it."
To your star rating, identical. To a human skimming, roughly the same. To an assistant answering "trusted BMW repair near me," the first is unusable and the second answers four separate questions: do they handle BMWs, are they honest about pricing, how fast are they, and what are they like to deal with.
The second review is not a slightly better review. It is a different category of asset.
Why this quietly changes what you should ask for
Most review requests are built to reduce friction: one tap, one click, done. Which is correct, and which is also why you end up with a wall of five-star reviews containing no information.
You are optimising for the wrong thing. A slightly lower response rate with substantially more detail beats a wall of empty stars, because the empty stars help your rating and nothing else.
How to ask for something usable
Ask a question, do not give a script
"What did you come in for today?" produces a review that names the service. "Would you mind mentioning our fast turnaround?" produces a review that sounds like you wrote it, because you did, and both Google and the reader can tell.
Ask while it is specific
A request that arrives an hour after the visit gets the detail. A request that arrives four days later gets "great service," because by then that is genuinely all they remember.
Name the person
"How did you find your appointment with Maria?" invites a review mentioning Maria. Reviews naming your staff read as real, they are excellent for staff morale, and they give an assistant a specific thing to say about you.
Let the unhappy ones through
Filtering so only happy customers reach Google violates Google's policies and sits badly with the FTC's rules on review suppression. It also produces a review profile that reads as suspiciously uniform. Offer everyone a private channel to raise a problem first, then let all of them review.
What this means for the reviews you already have
They still count. Volume and rating have not stopped mattering, and a long history of them signals an established business.
What has changed is the marginal value of the next one. If you already have three hundred reviews, review three hundred and one adds almost nothing to your rating. If it is the first one that says you handle a specific service that nobody has mentioned before, it might be the reason you get named for that question.
So stop counting. Start reading them the way a machine would, and ask yourself what questions they could answer. The gaps are your list.
Common questions
Is it against the rules to tell customers what to write?
Telling someone what to say crosses a line. Asking a question that prompts them to describe their own experience does not. “What did you come in for today?” is a prompt. “Please mention our same-day service” is a script, and it will read like one.
Does review length actually matter?
Specificity matters more than length. Two sentences naming the service, the outcome and the timeframe are worth more than a paragraph of adjectives, because those two sentences can be quoted as an answer to a real question.
See what your reviews are currently saying about you
The free report runs sentiment analysis across your existing reviews and shows what themes come up, which is usually the fastest way to find out what you are missing.
How AI visibility works → or get a free report on where you stand