Why reviews now decide who AI recommends

How AI assistants pick local businesses

When a homeowner in Patchogue asks ChatGPT, Gemini or Google’s AI Overviews for “a good roofer near me who handles storm damage,” the assistant has to choose a handful of names from many. It doesn’t have personal experience, so it relies on what it can find and summarize: business listings, websites, local directories, news mentions and, very often, reviews.

Reviews are useful to these systems because they are written by customers, they are plentiful, and they describe specific experiences. A business with lots of reviews mentioning fast storm repairs in Suffolk gives an AI system something concrete to say. A business with a few old one-line reviews gives it very little.

The exact way each assistant weighs sources isn’t public and changes over time. But the pattern is consistent enough to act on: the businesses that get recommended tend to have strong, recent, specific reviews.

What makes a review useful to AI

Star ratings matter, but the text matters more than many owners realise. A review that says “Great job” tells nobody much. A review that says “They replaced our boiler in Merrick the day after it failed and explained every option” tells people and AI systems the service, the location, the speed and the experience.

Recency matters too. A steady flow of reviews shows that the business is active and that recent customers are still happy. A burst of reviews from two years ago followed by silence raises questions.

Volume, consistency across platforms and how you respond all contribute. None of this is a formula, but together they form the picture AI systems are trying to read.

• Specific: names the service performed.
• Local: mentions the town or neighborhood naturally.
• Recent: part of a steady flow, not a one-off burst.
• Detailed: describes what happened and why it went well.
• Answered: has a thoughtful reply from the business.

How to ask for better reviews

You can’t write reviews for customers, and you shouldn’t script them. But you can make it easy and prompt people to include useful detail. The best time to ask is right after the job, while the customer is pleased and the details are fresh.

A short text with a direct link works well. You might say something like: “Thanks for choosing us for your AC repair. If you have a minute, a review mentioning what we did helps other neighbors find us.” That nudges people toward specifics without telling them what to say.

Make asking part of the routine rather than something you remember occasionally. Technicians can ask in person, the office can follow up by text, and a reminder can go out a few days later for anyone who hasn’t responded.

Where reviews should live

Google reviews are the priority for most Long Island businesses because they appear on your Business Profile and in Maps. But AI assistants look across many sources, so it helps to have a presence elsewhere too.

Depending on your industry that might be Yelp, Facebook, Angi, Houzz, Nextdoor, Healthgrades for dentists and chiropractors, or Avvo for law firms. You don’t need to chase every platform; pick the two or three that matter most in your field and keep them active.

Your own website matters as well. Featuring real reviews on service and area pages, with structured data where appropriate, ties customer feedback directly to the services and places you want to be known for.

Responding to reviews, including the bad ones

Replies are part of your reputation. A thoughtful response to a positive review reinforces what the customer said. A calm, helpful response to a negative one shows future customers how you handle problems, which can matter more than the complaint itself.

Keep replies specific and human. Thank the customer by name if appropriate, mention the service and avoid copy-pasted responses. For complaints, acknowledge the issue, avoid arguing in public, and offer to resolve it directly.

If a review breaks the platform’s rules, such as a fake review or one from someone who was never a customer, report it through the proper process. Don’t try to bury it with fake positives.

What not to do

Buying reviews, swapping reviews with other businesses, or offering discounts in exchange for five stars all break platform rules and, in some cases, consumer protection law. Platforms are increasingly good at spotting them, and removed reviews or suspended profiles can set you back a long way.

Review gating, where you only ask happy customers to post publicly, is also against Google’s policies. Ask everyone, and use the feedback from unhappy customers to fix what went wrong.

Finally, don’t ignore reviews once they arrive. An unanswered pile of feedback suggests nobody is paying attention.

Reviews as part of a bigger picture

Reviews are powerful, but they work best alongside a clear website, an accurate Google Business Profile and consistent listings. AI systems are trying to build a confident picture of who you are, what you do and where you do it. Reviews provide the proof; the rest provides the facts.

That is the approach we take with Long Island clients: make the business easy to understand, then make it easy to trust. We’ve seen how much structure and clarity can matter in our own products, such as What Is It Worth?, which earns more than 100,000 organic clicks a week. The same principles apply to a local contractor, just at a different scale.

No one can promise that an AI assistant will recommend you. What you can do is make sure that when it looks, it finds plenty of good, specific evidence.