45% of Americans now ask AI for a local business

Who the 45% actually describes

The Local Consumer Review Survey 2026 ran on a representative panel of 1,002 US adults. The AI chapter, published on 10 March 2026, narrows most of its cuts to the 455 respondents who had used an AI tool for a local business recommendation in the previous twelve months. Within that, the share of people who had done it at all climbed from 6% in 2025 to 45% in 2026, which makes AI the third most-used tool for local business recommendations, behind Google and Facebook and ahead of Yelp and Tripadvisor.

Two qualifications before you repeat it. The panel is American, so the figure describes the United States and nowhere else. If you have seen 45% quoted as a UK number, and we have watched an AI summary hand it out that way, that version is wrong. The 6% also comes from the previous year's wave, and BrightLocal does not publish the 2025 question wording, so read the jump as fast growth rather than a precise multiple.

The shape underneath the headline is the part worth keeping. Adults aged 30 to 44 lead at 64%, while the over-60s sit at 24%. Most of it happens in one product: ChatGPT at 31%, with Google's AI Mode second at 23%. If your customers are in their thirties and forties, this is already routine behaviour for them.

What 45% does not mean is that 45% of local searches now happen in a chatbot. BrightLocal's separate channel study puts AI as the first channel for 8% of US local searches and somewhere in the journey for 23%, against Google Search at 71% on that same journey basis. Of the people who did start on an AI tool, 43% went on to check Google and only 18% felt ready to contact the business without checking further. The AI answer builds the shortlist, and something else closes it.

Most of what gets cited is ordinary business websites

The more useful half of the story is on the supply side, in a study BrightLocal published in August. It ran close to two million citations across more than 116,000 domains through Google AI Overviews, Google AI Mode and ChatGPT in the US, UK and Australia. The finding a small business should hear first: around 93% of the domains cited were individual business websites. Those are sites belonging to one business, rather than aggregators, listicles or the big platforms.

The largest single source is the one you already own. Google Business Profile accounts for 28.5% of all citations in the set. Yelp is the other headline: it appeared as a source in 80% of ChatGPT's local answers. Before an Irish reader goes off to claim a Yelp page, that number is a US artefact. Yelp is thin here and weak in Britain, and the study publishes no country split, so it cannot tell you which directories win in Cork. Google Business Profile is the finding that transfers.

BrightLocal describes an AI local answer as assembled from reviews, local directories, business websites, social media and other third-party sources, varying by tool and by prompt. In practice, for a hotel in Galway or a dentist in Kilkenny, that resolves to a short and boring list:

• Your Google Business Profile. The name, category, address, opening hours and photos on that profile are the single biggest input, and the one you can correct this afternoon.
• Your own website. Most cited domains are business sites, so the facts have to be on your pages in readable text, not locked in a booking widget or a PDF menu.
• Whatever review profile you actually have. Reviews are both a source of sentiment and, as the next section shows, a condition of being mentioned at all.
• The directories your market uses. In Ireland that means Tripadvisor for hospitality and the trade or professional bodies for regulated services, rather than the US directory mix in the study.
• Social profiles and third-party mentions. Anything else about you the engine can reach, including the press coverage and listings you did not commission.

There is no Irish version of this number

We looked. No published dataset measures how many people in Ireland use AI tools to find a local business, and nobody should pretend the US 45% stands in for one. What exists is a set of adjacent Irish figures pointing in the same direction.

Friday analysed 175 million Irish searches and found click-through rate fell 26% year on year, from 3.62% to 2.67%, while average position barely moved, 6.38 to 6.18. Same rankings, a quarter fewer clicks. The CSO reported in August that more than 80% of Irish 18 to 29-year-olds understand AI well enough to explain it to someone else, third highest in Europe. The Reuters Institute has 7% of Irish news consumers using AI chatbots for news against 4% in the UK. And a 2025 four-country survey covering the US, UK, France and Germany put 19% of consumers using AI tools to find local businesses, with almost one in four Americans doing it against fewer than one in five across all four countries. Ireland sits behind the US on this, then, but on the same curve, and we have already found Irish holidaymakers planning trips with AI.

The figure that should bother you is about what happens after the answer appears. Across markets, only 4% of people say they always or often click through to the underlying sources of an AI answer, against 19% from search. In the UK it is 1%. For most readers the summary is the entire encounter with your business, and you never see the impression it left.

Reviews decide whether you are in the answer at all

Seer Interactive analysed 804,491 AI responses across 1,926 brands and found that brands with no verified active review profile had a median citation rate of 1%, while brands with even a minimal profile were cited 53.5% of the time. It was measured on Trustpilot and spans B2B as well as consumer brands, so the honest reading is an active third-party review profile rather than Google reviews specifically. We went through the study and its funding in a separate piece.

The same study found 99.5% of AI citations arrived by way of organic search rankings, which is a quiet argument against treating AI visibility as a separate discipline with its own budget line. Ranking still feeds the answer.

And the answer is short. In the US, 72% of consumers look at three or fewer businesses before deciding. When we followed the citations behind two Dublin hotel questions ourselves, fourteen of the nineteen sources were Google Maps entries and five were hotel-owned pages. Nothing in that run came from a review site or a travel magazine. One town, one category, so treat it as an illustration rather than a law. It does match the citation study: the engine reaches for the profile and the business's own page.

Trust in what it produces is higher than the sceptics expect and lower than the vendors claim. Among people who use these tools, 64% trust AI as much as reviews for local recommendations and 63% trust the recommendations themselves. They still check: 97% at least sometimes double-check against real reviews and 88% look at whether a review is legitimate or where it came from. On the February report, 82% of consumers read AI-generated review summaries and 23% would be happy to decide on the summary alone. Those are US figures too, and they are about a paragraph of text describing your business that you have almost certainly never read.

Five things, in this order

For a hotel, a dental practice, a garage or a restaurant in an Irish or UK town:

• Finish the Google Business Profile. Category, address, opening hours, phone, service list, photos taken this year. It is 28.5% of citations in the biggest study we have, and most profiles are half done.
• Put the facts on your own site in plain text. Address, hours, prices or price ranges, what you actually do and do not do, the areas you cover. If a fact only exists inside a booking iframe or a PDF, treat it as absent. Business websites are 93% of cited domains, so this is where the engine goes looking.
• Make the profile and the site agree. Two different opening hours, two spellings of the business name or an old address in one place all become the reader's confusion, and the engine has no way to referee it.
• Ask for reviews, and answer them. A profile with recent reviews and visible replies is the difference between 1% and 53.5% in Seer's data. Answer the bad ones properly, because the reply is part of what gets summarised.
• Go and read what the engines say about your category in your town. Ask ChatGPT, Gemini and Google AI Mode for the best option in your category in your town, and write down who they name and what they say about you. Do it more than once, on different days.

That last one is the step people skip, and it is the only one that tells you whether the first four worked. If you would rather not sit there re-typing the same prompt into three tabs every fortnight, that is what a mention scan does: it runs your questions across the engines on a schedule and keeps every answer, so you can see what changed and which sources it came from.

See what AI says about your business

Questions people ask about this