We Scanned 50 Local Service Sites: Here's What AI Can't Find

Last month, we ran a simple experiment. We picked 50 local service websites, roofing, HVAC, plumbing, landscaping, and garage door companies across the US. and scanned each one against ChatGPT, Perplexity, and Google AI.

The results were worse than we expected.

The Bottom Line

86% of local service sites scored below 30/100 on AI visibility. Only 2 sites were cited by all three AI engines. The rest? Invisible.

What We Measured

For each site, we checked:

  • Does AI mention the business by name?
  • Does AI know their service area?
  • Does AI cite their website URL?
  • Is their pricing or offers referenced?
  • Are they recommended over competitors?

What AI Can't Find on Most Sites

1. No Structured Service Data

78% of sites had no LocalBusiness schema. Without structured data, AI reads raw HTML and often misinterprets what the business actually does.

A site listing "residential and commercial services" gets parsed as vague. A site with schema saying "Roofing Contractor" + "Houston, TX" gets cited immediately.

2. City Pages Are Missing or Thin

64% of sites had only a homepage and contact page. AI needs city-specific pages to associate a business with a location.

Homeowners don't search "roofer." They search "roofer in Katy" or "emergency roof repair Cypress." If you don't have pages for those queries, AI can't cite you.

3. GBP and Website Don't Match

57% had mismatched name, address, or phone between Google Business Profile and the website. AI cross-references both. When they don't match, trust drops. and citations disappear.

4. Reviews Aren't Helping

51% had fewer than 10 reviews visible to AI crawlers. Even when reviews exist, they're often hidden behind JavaScript or login walls that AI can't read.

AI reads review content for specific signals: punctuality, cleanup, fair pricing, professionalism. Generic 5-star reviews without detail don't move the needle.

The Winners Shared 3 Traits

The 2 sites that were cited by all three engines shared three things:

  1. City-specific pagesfor every town they served
  2. LocalBusiness schemawith accurate service areas
  3. Review snippetson the homepage with specific homeowner language

One was a small HVAC company in Phoenix with 23 reviews. The other was a garage door repair shop in Orlando with 41 reviews. Neither had a big marketing budget. They just had the right technical setup.

What This Means for Local Trades

The gap between "invisible" and "cited" is much smaller than most business owners think. It's not about content volume. It's about structured data, location pages, and review visibility.

AI doesn't need long blog posts. It needs clear, consistent, structured signals that say:"This is a real business, here's exactly what they do, here's where, and here's proof people trust them."

The Quick Audit

Check your own site in 5 minutes:

  1. Run your domain through theSottoVox scanner
  2. Look for schema errors in Google Rich Results Test
  3. Search your city + service. do you have a page for it?
  4. Compare your GBP and website NAP. do they match exactly?
  5. Check if AI can read your reviews

TL;DR

We scanned 50 local service sites. 86% were invisible to AI. The missing pieces were schema, city pages, GBP consistency, and visible reviews. The fix is straightforward. and smaller companies are already winning with it.

Scan your site free →