Original research · October 2026

AI Can Read Most Trades Websites. It Can't Tell What Most of Them Are.

· Published October 13, 2026

We audited the homepages of 268 home-service and trades businesses in 12 US metros the way AI crawlers see them. Only 1% block AI crawlers. But 28% don't describe the business in structured data, and 84% have no marked-up answers an assistant could quote. For most trades businesses, the gap isn't access. It's identity and answers.

Key findings

  • Only 1.1% of home-service websites block AI crawlers in robots.txt.
  • 6.7% show almost no text without JavaScript, so AI crawlers see a near-empty page.
  • 27.6% have no LocalBusiness or Organization schema telling AI what the business is.
  • 83.6% have no FAQ or HowTo markup, the format assistants find easiest to quote.
  • The median AI-readiness score was 75/100; 19% scored under 50.

How often each check failed

Share of the 268 audited homepages failing each check, highest first.

No FAQ or HowTo schema83.6%
No llms.txt66.8%
Under 70% of images have alt text41.4%
Meta description missing or wrong length38.8%
Not exactly one H1 heading36.6%
Title tag missing or wrong length30.6%
No LocalBusiness or Organization schema27.6%
No structured data at all19.8%
Very little text without JavaScript6.7%
No XML sitemap6.3%
robots.txt blocks AI crawlers1.1%

Download the aggregate data: ai-search-readiness-2026.csv (CC BY 4.0, please credit InfuseAI Agency).

What it means for a trades business

The common fear is that AI can't see small business websites. For trades businesses, that's mostly not true: almost every site we checked lets AI crawlers in, and most show real text without JavaScript.

The problem is what AI finds when it gets there. More than a quarter of sites never state, in a form a machine can read, what the business is, what it does and where it works. Most offer no direct, marked-up answers to the questions homeowners ask. An assistant choosing two or three companies to name will pick the ones it can describe confidently. It gets that confidence from the site itself and from other sources that agree with it.

Three fixes cover most of the gap: add LocalBusiness schema with your real name, phone and service area; publish short, answer-first pages for each service; and make sure directories, reviews and local sources describe you the same way.

Method

  • Sample: named, non-chain businesses in OpenStreetMap tagged as plumbers, electricians, HVAC, roofers, general contractors, landscapers or painters that list a website, in Phoenix, Dallas, Chicago, Seattle, Charlotte, Tampa, Kansas City, Columbus, Nashville, Salt Lake City, Denver and Minneapolis. Up to 60 per metro, chosen at random. 319 websites were listed; 268 loaded and were audited.
  • What we checked: each homepage, its robots.txt and its XML sitemap, fetched without running JavaScript, the way most AI crawlers read a page. We used the same automated audit we run on client sites. No business was contacted.
  • Limits: homepages only. OpenStreetMap over-represents established businesses. Robots.txt rules don't show firewall-level bot blocking, so 1.1% is a minimum. 268 sites is a snapshot, not a census, and samples per trade or city are too small to compare.
  • Dates: October 12–13, 2026.

Want to know how your own site scores? Run the free AI Visibility Check. It uses the same checks.

See Where Your Site Stands

Check your homepage in 20 seconds, or book a call and we'll show you who ChatGPT recommends instead of you.