Half your customers now ask AI before they call. Here is the fix.

The number is the news, then the fix

If your phone has been quieter this spring and Google still shows your rankings holding, the reason might be a place on your Google profile you haven't checked in a while. AI use for local business recommendations climbed from 6 percent of US consumers in 2025 to 45 percent in 2026, per the BrightLocal Local Consumer Review Survey released in May 2026. A year ago that was one in seventeen customers; today it is nearly half. That's the largest 12-month shift in customer search behavior in the last decade.

The number that goes with it is the one to act on. 74 percent of consumers said they only weight reviews from the last 90 days when picking a local business. That's nearly three in four. ChatGPT and Gemini apply roughly the same filter when they read your profile to pick who to recommend. A 4.7-star Google profile from a practice that stopped collecting reviews in 2024 still looks acceptable when a human reads it. The AI tool reads the silence and picks the practice on the same block with 14 fresh reviews from the last 60 days.

When I say AI citation, I mean the businesses ChatGPT and Gemini actually name when someone asks for a local recommendation.

What the AI tool actually reads

When a customer asks ChatGPT or Gemini for a local business, the tool reads a handful of pages it has already decided are relevant rather than crawling the web in real time. For local recommendations, that handful is dominated by your Google Business Profile (your GBP), the recent Google reviews on it, your star average, and a few supporting pages from Yelp and your own site.

The AI tool is hunting for the facts a customer would want: the most recent reviews, the current star average, service names, hours, whether the business has been responding to reviews, and photos with dates. The tool weights these signals like a sharp human shopper would, then it weights them harder. Sit silent for 90 days and your profile reads as quiet to the AI, which then picks the practice down the street with eight reviews from the last six weeks at 4.9 stars.

The same BrightLocal study found that 31 percent of consumers now refuse to use a business below 4.5 stars on Google. That number was about half that two years ago. Customers and AI tools both got pickier in the last two years.

Why your old review profile isn't the asset it was

Three years ago, getting to 200 Google reviews and a 4.6 star average was a finished job. The number sat there as a credibility marker for the next person who scrolled. Today the reviews from 2022 do almost no work for an AI recommendation tool. The tool is reading the recent set and skipping the old set.

I run dozens of these audits a quarter, and this is the gap I see most often. Take the practice with a 4.8 star average from 230 reviews, with the most recent review dated October 2024. Google search shows the practice in position three locally. ChatGPT, asked for a recommendation in the same area, names two competitors and skips the practice. The competitors named have 80 to 120 reviews each, but 18 of them landed in the last 90 days.

Now take a practice across town that did the opposite. 4.6 star average. 95 reviews total. 22 of them in the last 60 days. Google search shows the practice in position five locally. ChatGPT names this practice first. Gemini agrees. The recent volume and recent rating did the work that the older, deeper profile could not.

The 30-day plan

Pick one Saturday afternoon, today if you can, and run this sequence.

1. Open your Google Business Profile on a phone. Look at the date of your most recent review. If it's more than 30 days old, the next four weeks are about closing that gap.
2. Make a list of 12 to 20 happy customers from the last 60 days. Recent visits are warmer asks than visits from last year.
3. Send each one a short, plain text message the day after their visit. Two sentences about the visit and the ask, plus a direct link to your Google profile. No discount attached. A plain text from you wins against a templated marketing email with a discount attached by a wide margin.
4. Reply to every existing review on your profile that you haven't yet replied to, in plain language without a script. Two sentences each is fine. Ten reviews replied to in twenty minutes. The reply is read by future customers and by the AI tool.
5. Hold the pace at three new review requests per week for the rest of the 30 days. Put it on the calendar as a recurring 20-minute block on Saturday morning.

The expected result, based on what I see in audits when an operator runs this exact sequence, is 12 to 20 fresh reviews in 30 days. Most of them at 5 stars. Average rises a tenth or two if the existing set was below 4.7. ChatGPT and Gemini start updating who they recommend at the four to six week mark. A second BrightLocal segment on AI trust found that 42 percent of consumers already trust AI recommendations as much as traditional reviews, and 63 percent of active AI users do. The result of the 30-day push lands in front of those customers directly.

Past day 30 and the rest of the work

The whole sequence runs through a phone, a list of recent customers, and the Google Business Profile dashboard. The cost is your time on Saturday afternoon and your front-desk staff's time during the week.

The second piece is staying past day 30. If you stop asking after the sprint, you stop showing up in AI recommendations within three months. A practice that runs the 30-day push and then goes silent will land back there within a quarter. Make the three-per-week ask a recurring habit on a calendar slot.

The review work is also one piece of a bigger job. Service-area pages (the pages on your site for each town or service zone you cover), plain-text pricing, and the Google Business Profile fields all matter. Recent reviews are the change that moves fastest on AI recommendations and the only one most operators can start today without a developer or a new vendor.

Where to begin this weekend

Open your Google Business Profile. Look at the date of the most recent review. If it's older than 30 days, you have your starting point. Send three short messages today to recent happy customers asking for a review. Put a Saturday morning calendar slot for the next four weeks to do the same. That's the work.

Frequently asked questions

Do I need to ask every customer for a Google review or is that pushy?

Asking three or four happy customers a week is normal practice across service businesses. The customers who had a good visit expect to be asked. The pushy version is a five-message follow-up sequence with a discount attached. The healthy version is one short text the day after the visit with a direct link to your Google profile. Most operators find that asking explicitly raises the response rate from about 5 percent to roughly 30 percent on warm visits.

What if my recent reviews are mixed and the average drops below 4.5 stars?

Keep asking. Address a 4.3-star average by adding 20 honest 5-star reviews from happy customers over the next 60 days. The math works in your favor when the recent set is good. AI tools and human readers both weight the last 90 days much harder than older reviews. Second, reply to every negative review in plain language without a script. The reply is read by future customers and by the AI tool building the recommendation.

How long until my review work shows up in ChatGPT and Gemini?

Most AI tools update who they recommend every two to four weeks. A 30-day push that adds 12 to 20 fresh, high-rated reviews to your Google Business Profile typically starts to move recommendations by week six. Your reviews keep moving recommendations as long as you keep asking. The pattern that fails is a one-month sprint followed by silence, because if you stop asking, you stop showing up in AI recommendations within three months.

Does the same review profile work across ChatGPT, Gemini, and Google AI Overview?

Mostly yes. All three lean heavily on Google reviews because that's where most of the local-business information lives. Each tool gets to the same data by a slightly different path: ChatGPT through Bing plus Yelp and a few directories, Gemini straight from Google's own profile fields, Google AI Overview built on top of Google's own data. There are only a few small differences. The main thing all three look at is your last 90 days of Google reviews.