The question "where's the best plumber / dentist / restaurant near me?" no longer goes to Google alone: it goes to ChatGPT and Gemini. And the AI doesn't hand back ten links to choose from — it recommends one, two, three. The number almost nobody has clocked yet: in BrightLocal's 2026 local consumer survey, 45% of consumers used AI to find local businesses in the past year, up from 6% a year earlier. It's now the third local discovery channel, behind only Google and Facebook. This guide is the map to being the business the AI names.
For a local business, AI is already the third discovery channel
If you sell to your neighborhood, your city, your region, this shift hits you first. Local queries are the most specific and highest buying-intent there is: "near me," "open now," "that does X" are people about to spend, not browsing. And more and more of them start that journey in an assistant, not on a map. The jump from 6% to 45% in twelve months that BrightLocal measures isn't an early-adopter curiosity: it's local buying behavior moving while most businesses look the other way.
That's the opening. Unlike ecommerce, where you compete against global catalogs, in local you compete against the four or five businesses in your area — and almost none of them are working this. Whoever cleans up their data and reviews today shows up in the answers while the place next door still thinks AI "is for the big guys." The window is open, and it won't stay open forever.
Each assistant drinks from a different source: Google AI from your Business Profile, ChatGPT from Bing
The number-one mistake in local is treating every assistant as if it were the same. They aren't, and the difference changes your to-do list. When Gemini or Google's AI Overviews answer a local query, they read your Google Business Profile (formerly Google My Business) as the authoritative source: it is, literally, the spine of the answer. But when ChatGPT browses the web to answer something local, it leans heavily on the Bing index and Bing Places — not Google.
Translated to your work, that's two listings, not one:
- Google Business Profile, complete and alive: it's what feeds Gemini, the AI Overviews, and the classic map pack.
- Bing Places, claimed and complete: it's the deck ChatGPT deals from when it answers local queries by browsing the web.
- Your industry directories (plus Apple Maps, Yelp, and whatever rules your niche): they're the backup sources the assistant uses to confirm you exist and your data lines up.
Your Google Business Profile is the spine: complete it and keep it alive
If local AI leans on your Business Profile, that profile isn't a one-time form you fill out: it's the asset you guard most. And "complete" isn't "created": it's the right primary category, services and attributes in detail, real hours (holidays included), service area, recent photos and, above all, activity. Profiles untouched for more than 30 days lose impressions; AI, like Google, rewards the "this is still alive" signal.
With the listing complete, the rest of the profile work is four concrete things:
- Pick the category and attributes well. They're the filters the assistant uses to narrow "best X that does Y near me." The generic category doesn't catch the specific query; the missing attribute is a filter you don't show up in.
- Post regularly. Updates, offers, Q&A: each bit of activity is an alive-profile signal and a chance to hold the exact words people ask.
- Answer the real questions on the listing itself. The "people also ask" of your Profile is direct raw material for the AI's answer: answer it yourself, in your words, before a third party does.
- Keep hours and data current. A wrong "open now" doesn't just annoy the customer: it teaches the assistant your listing isn't reliable — and reliability is what decides who gets named.
Reviews and data consistency: what gets you in or leaves you out
Here's the heart of the local game. In the industry counts, AI leans on your website, your directory listings and your reviews in around 86% of its local recommendations; the weight falls mostly on your website content (about 24%), reviews (around 16%) and the accuracy of your directory listings (near 13%). It's not one magic button: it's that all three legs say the same thing and tell the truth.
And the leg most businesses flunk isn't glamorous: data consistency, what the trade calls NAP (name, address, phone). If your business shows up as "Pérez Clinic," "Pérez Dental Clinic" and "P. Dental Clinic LLC" depending on the site, the AI can't tell they're the same and, in doubt, drops you. NAP inconsistency across platforms is one of the most common reasons a local business gets left out of the answer. Four things, in order:
- Unify NAP everywhere. Same name, same address, same phone — character for character — on your site, Google, Bing and every directory. It's boring work and it's the one that moves the most.
- Put reviews to work and respond. Recent reviews with an owner reply weigh double: they feed the summary the AI writes about you and prove you're alive. Ask for them systematically and reply to the bad ones too: a credible profile gets cited more than a perfect, suspicious one.
- Mark up your site with LocalBusiness schema. Structured data with your name, address, geocoordinates, hours and price range lets the AI read your listing without guessing. Geocoordinates in the schema are one of the strongest signals for a local AI Overview to cite you.
- Write on your site what customers actually ask. A page per service and per area, answering in their words ("do you do Sunday emergencies in Lavapiés?"), is citable content. The generic homepage with a phone number gives the AI nothing to recommend.
Measure it like a map, not a list: do you show up for "near me" searches?
Local GEO measurement isn't counting mentions in the abstract: it's checking whether you show up when someone in your area asks with intent to go. Define 15 to 25 real local queries for your category — the way people actually phrase them: "best X near me," "X in [your neighborhood] open now," "X that does Y in [your city]" — run them once a month in ChatGPT, Gemini and Perplexity, and note per query whether you appear, in what spot, and with what data (phone right? hours right?). The month-over-month trend is what counts, not a single day's snapshot.
And check the boring stuff, where people fall down: that your Business Profile is complete and updated this month, that Bing Places exists and matches Google, that your NAP is identical across the main directories, that your LocalBusiness schema validates without warnings, and that your site doesn't block the crawlers that fetch pages to answer. On timing, be honest with yourself: first citations can land in 3-5 days after refreshing the listing and the site, but consistent visibility across platforms takes 2 to 4 months of sustained work.