Local businesses are in a strange position with AI search. On one hand, the queries are exactly what AI assistants are best at — "best dentist in [city]," "good roofer near me," "Italian restaurant open late in [neighborhood]." On the other hand, most local businesses have done zero optimization for these queries. Their websites are built for foot traffic and phone calls, not for being parsed by an AI assistant. The result: when someone asks ChatGPT, the AI picks two or three businesses to recommend, and most local businesses aren't in that shortlist.

We've audited a lot of local business sites over the past six months. The patterns are consistent. This is the exact playbook for local service businesses — plumbers, dentists, roofers, restaurants, salons, contractors — who want to be the recommendation, not the also-ran.

The unique stakes for local businesses

When ChatGPT answers a SaaS query, it might list three or four products as recommendations. When ChatGPT answers a local query, it often returns just one or two business names. The funnel for local services is brutally compressed. Either you're in the AI's top recommendations, or you're invisible.

Worse: local search queries on ChatGPT (and even more so on Google AI Overviews) are growing fast as users get used to asking the AI for nearby recommendations. Some categories — restaurants, home services, healthcare — already see 20–30% of discovery happening through AI assistants. The number is going up.

Local SEO best practices still matter (Google Business Profile, reviews, location pages), but they aren't enough on their own. AI assistants weight different signals than Google Maps does. Here's what they actually use.

What AI assistants look for in local recommendations

When ChatGPT answers a local query, it weighs roughly these signals:

  1. LocalBusiness schema — does the site have explicit JSON-LD markup declaring it's a business, what services it offers, what its address is, and what hours it operates?
  2. Geographic anchoring in content — does the site clearly state where it operates (city, neighborhood, service area)?
  3. Service specificity — does the site name the specific services offered, not just generic categories?
  4. Review signals across the web — Google, Yelp, industry directories, news mentions.
  5. Authority signals — being mentioned in local news, "best of" lists, community blogs.
  6. Operational signals — current hours, current phone number, current address. Stale info gets penalized.

The first three are entirely within your control. The last three take longer. Start with what you control.

Fix 1: Add LocalBusiness schema (the highest-leverage 30 minutes you'll spend)

If you run a local service business and you have nothing else from this list, do this. LocalBusiness schema is JSON-LD code that explicitly tells AI assistants: "I am a [type] business, located at [address], serving [area], offering [services], open [hours]." Without it, AI engines have to guess everything from your prose, and they usually guess wrong.

Here's a minimal LocalBusiness schema you can adapt:

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Plumber",
  "name": "Acme Plumbing",
  "address": {
    "@type": "PostalAddress",
    "streetAddress": "123 Main St",
    "addressLocality": "Vancouver",
    "addressRegion": "BC",
    "postalCode": "V6B 1A1",
    "addressCountry": "CA"
  },
  "telephone": "+1-604-555-0100",
  "url": "https://acmeplumbing.com",
  "openingHours": "Mo-Fr 08:00-18:00",
  "areaServed": ["Vancouver", "Burnaby", "Richmond"],
  "priceRange": "$$"
}
</script>

Replace the values with your real business info. Pick the most specific schema type that applies (Plumber, Dentist, Restaurant, Roofing, BeautySalon, etc — schema.org has hundreds). Drop this into your homepage <head>. Test it with Google's Rich Results Test. Done.

Fix 2: Anchor your content geographically

Local AI queries always include a place — "in Vancouver," "near me," "in the Mission district." For the AI to surface you, your content needs to clearly state where you operate. Most local business sites are surprisingly vague about this. They have an address in the footer and no other geographic anchoring.

Where to add geographic anchoring:

  • Homepage hero: your headline or sub-headline should mention your city or service area. "Vancouver's most-booked emergency plumber" beats "Fast, reliable plumbing."
  • Service pages: if you have a "Drain Cleaning" page, the H1 should be "Drain Cleaning in Vancouver, BC" — not just "Drain Cleaning."
  • Neighborhood pages: for businesses serving multiple areas, build a page per neighborhood. "Plumbing services in Kitsilano," "Plumbing services in Mount Pleasant." Each page is real content (not duplicate), but they each anchor to a specific area.
  • About page: the more specific your "where we are and who we serve" content, the better.

Fix 3: Name your services specifically

AI assistants struggle with vague service descriptions. "We provide comprehensive plumbing solutions" tells the AI almost nothing. "We do drain cleaning, water heater repair, leak detection, sewer line replacement, and emergency plumbing" tells the AI exactly what to surface you for.

Audit your service descriptions. For every service you offer, ask: would someone asking ChatGPT for that specific service find a match on my page? If your roof repair page doesn't include the phrases "leak repair," "shingle replacement," "flashing repair," and "emergency roof tarping" — none of which most homeowners would type, but all of which they'd describe to an AI — you're missing matches.

Fix 4: Build review signals across the web

AI assistants weight third-party verification heavily for local businesses. The single review on your own site says nothing. Reviews on Google Business Profile, Yelp, Angi, HomeAdvisor, Trustpilot, BBB, industry-specific platforms — these collectively tell the AI "this business is real, has customers, and gets results."

If you have fewer than 30 Google reviews, that's your priority. Most local businesses leave dozens of unfilled review requests on the table. Build a system: every completed job, every satisfied customer gets a polite review request. Within 60 days you can move from 8 reviews to 50, and the AI citation impact is significant.

Fix 5: Get listed in local directories and "best of" content

When an AI assistant is deciding which businesses to recommend in a local query, it looks at how often the business appears across the web. A business listed in three directories and mentioned in one local news article is in a completely different league from a business with only its own website.

Tactical wins, in priority order:

  • Google Business Profile (free, mandatory)
  • Yelp listing (free, takes 10 minutes)
  • Bing Places (free, often forgotten — but matters since ChatGPT uses Bing)
  • Apple Maps Connect (free)
  • 3–5 industry-specific directories (BBB, Houzz for home services, Healthgrades for medical, etc.)
  • Local chamber of commerce
  • Local "best of" blog posts (pitch yourself to bloggers writing year-end roundups)

Each listing takes 10–20 minutes. Doing all of them is a 4–6 hour project that pays off for years.

The reality about local AI search

One thing we routinely tell local business owners: you probably can't beat the chains on brand recognition. ChatGPT will name Home Depot, Walmart, McDonald's, and major franchises by default — they have overwhelming training-data exposure.

What you can do is win the queries the chains don't dominate. "Best independent plumber in [neighborhood]," "family-owned roofer in [city]," "small business [service] near [area]" — these are queries where the local mom-and-pop has a real shot. The customer asking these questions is already filtering against chains. Technical readiness + reviews + directory listings can put you in front of them.

Run our free audit on your business site to see exactly where you sit. We routinely find local businesses with no LocalBusiness schema, no FAQ blocks, and blocked AI crawlers — three fixes that take less than 2 hours combined. The improvement window is typically 30–60 days after fixing. Some of the easiest AI search wins available are sitting in local business audits.

If your audit flags blocked AI crawlers, our 2026 robots.txt setup guide has the copy-paste configuration. If you're not sure why ChatGPT isn't surfacing you despite a clean setup, the eight-point diagnostic checklist covers what to investigate next. And for the bigger-picture explanation of how AI assistants weigh local-versus-national businesses, see how ChatGPT decides which sites to recommend.