The short answer: Google's AI now reads the text of your reviews, not just the star average — so the specific words customers use affect which searches you turn up for. A review saying "replaced our hot water tank in Brocklehurst, same day" makes you relevant to "emergency hot water tank Kamloops" in a way that five stars and no words never will. Ask in a way that prompts the detail (what job, where, what the outcome was), keep the flow steady rather than doing one big push, and never write or incentivise a review — that's against Google's policies and it's the fastest way to lose the ones you have. The review link and QR tool makes the asking a ten-second habit.
Most advice about Google reviews is really advice about arithmetic: get more of them, keep the average above 4.5, respond to the bad ones. All still true. But it treats a review as a number with some text attached, and in 2026 that's backwards. The text is the part doing the work.
The 2026 local-search research is fairly direct about this: Google's AI reads the content of reviews, not just the rating, which means the words your customers choose affect your relevance for related searches. Same for the AI assistants — when ChatGPT or Gemini decides which two businesses to name, it's reading language, because language is what it can actually reason about.
That's a small shift with a large practical consequence. It means two businesses with identical 4.8 averages can have completely different visibility, depending on whether their reviews say anything.
What a "good review" is now
Compare these two, both five stars:
"Great service, highly recommend."
"Our furnace died on the coldest night of the year and they had someone out to our place in Westsyde by 8am. Replaced it the next day, explained the options, price was exactly what was quoted."
The first one moves your average. The second one moves your average and tells Google — and every AI assistant — that you do emergency furnace work, that you serve Westsyde, that you replace units rather than only servicing them, that you quote accurately, and that you respond fast. Every one of those is a query someone is typing.
So the useful mental model: a review is a customer describing your business in their own words, in public, to a machine that is deciding who to recommend. Star ratings sort you. Review text is what makes you findable for the specific thing you actually do.
How to ask so people say something
You can't script it — and shouldn't try, for reasons I'll get to. But how you ask heavily shapes what you get, and almost everyone asks in the way most likely to produce "great service".
Ask right after the moment worked. The finished install, the fixed problem, the "wow, that looks great" on the walkthrough. That's when the specifics are still in their head. A request sent a week later gets a generic review because a week later it's a generic memory.
Ask a question instead of asking for a review. "Would you mind leaving a Google review?" produces one line. "Would you mind leaving a Google review — the thing that helps most is if you mention what we did and roughly where you are" produces the useful kind. You're not telling them what to say; you're telling them what's helpful, which people are generally happy to hear.
Make the mechanics trivial. Every extra step loses people. A direct link that opens the review box, sent by text while you're still in the driveway, converts far better than "search for us on Google". That's the whole reason I built the review link and QR code generator — paste your profile's share link, get the direct link, a QR code for the counter or the invoice, and copy-paste ask-scripts.
Ask everyone, steadily, forever. This is the boring one that decides everything. Ten reviews spread over ten months beats fifteen in one week and then nothing, because recency is now weighted heavily — a steady flow reads as an active business, a big old pile reads as a business that used to be busy. It also protects you: February 2026 saw a broad wave of Google review removals across profiles, and businesses treating their review count as a permanent asset they'd already banked took the hit hardest. A habit is durable; a campaign isn't.
What not to do
Two lines that are not worth crossing, and I'd rather be blunt than diplomatic about them.
Don't write reviews, or draft them for customers to paste. Beyond being against Google's policies, it's obvious. Reviews that all share a vocabulary and sentence rhythm read as manufactured to a human and pattern-match trivially for a machine. The value of review text comes precisely from it being independent — remove the independence and you've removed the signal you were trying to build.
Don't offer anything in exchange. No discounts, draws, or free coffee for reviews. It's a policy violation, and gated or incentivised reviews are exactly the kind of thing that gets swept in a removal wave. You can absolutely thank someone; you can't pay them.
The honest version works better anyway: "If we did a good job, a Google review genuinely helps a small local business — and if we didn't, tell me first and I'll fix it."
Respond in words that matter too
Your replies are also text on the profile, and they're the one part you control. Which means a reply is a second chance to state plainly what happened: "Thanks Dave — glad we could get the hot water back on same-day out in Barnhartvale." Natural, useful, and it reinforces the service and area without you keyword-stuffing anything.
Reply to the negative ones the same way: calm, brief, factual, and offline as fast as possible. If that's the part you dread, the review response generator gives you a template for each situation, and how to respond to a negative Google review walks through the formula.
The number you're aiming at
None of this replaces the arithmetic — you still need enough reviews and a healthy average to be taken seriously. If you want to know exactly how many five-star reviews it takes to get from where you are to 4.5 or 4.7, the star rating calculator solves it, including how the number changes as your total grows.
But once you're in the range, extra reviews with nothing in them stop adding much. Reviews that say what you did and where keep compounding, because each one adds another true sentence about your business to the pile that Google's AI and every assistant reads when deciding who to name. If you want the broader picture of how that decision gets made, Google's AI local pack shows 68% fewer businesses is the wider story, and how to get more Google reviews covers building the asking habit into your workflow.
And if reviews are one part of a bigger visibility problem, local SEO for Kamloops businesses is the whole job handled — profile, reviews, listings, and the content that makes you quotable.