24 US, Canadian and Australian sites, scored for AI search
I keep hearing that weak AI visibility is a developing-market thing, that businesses in the US, Canada and Australia have it handled. So I ran the same scan on 24 of them across law, dental, hotels, restaurants and med spas. Half scored an A. And the ones without structured data still sat 30 points below the ones with it. The gap is not about the market. It is about legibility.
The 24 sites are real independent businesses: 13 in the US, 6 in Canada, 5 in Australia. I scored each homepage once, on 17 July 2026, with the same free Report Card I use on Indian sites. Same engine, same signals, no JavaScript executed, because the major AI crawlers do not execute it either.
The market is not the problem
Twelve of the 24 scored an A, a much healthier median than the Indian samples I have published. But split them the same way, by whether the site publishes structured data, and the same fault line appears. The 18 with structured data have a median of 88.5. The 6 without have a median of 58.5. A 30-point gap, in three of the most developed digital markets on earth. Being in the US does not make a page readable. Adding the markup does.
Dental lagged, med spas led
By industry, the pattern was consistent with everything else I have scanned. Median scores:
- Med spas: 98. The most consistently well-built sites in the sample.
- Boutique hotels: 88. Strong, though the booking-widget problem still hides inventory from crawlers.
- Law firms: 87. Good structure, but many bury attorney credentials in prose instead of Person schema.
- Restaurants: 74. Menus as images and JavaScript widgets keep the actual food invisible.
- Dental: 70.5. The weakest group, same as in India: excellent practices, thin markup.
The samples per industry are small, so read them as hints. But the ordering matches what I see everywhere: the businesses whose information lives in images, PDFs and JavaScript score lowest, regardless of country.
Nobody was doing FAQ markup
Twenty-one of the 24 had no FAQ markup, the same near-total absence I found scanning Indian hotels and clinics. Across three countries and five industries, the single most quotable format for an answer engine is the one almost nobody uses. It is not a coincidence that it is also the cheapest to add: the answers are already written on these sites. The work is labelling them so a machine can lift them.
What this means if you are competing
Half of these businesses are already legible to AI. That cuts both ways. If your competitors are in the readable half and you are not, an answer engine has a reason to name them and not you, and it has nothing to do with which of you is better at the actual work. The fixes are the same everywhere: structured data first, then FAQ markup, then getting your real content out of images and JavaScript and into HTML text. The structured data guide walks through the first and highest-impact one.
Method and limits
24 homepages, fetched once each on 17 July 2026, scored with no JavaScript executed, homepage only. Per-country and per-industry cuts are small. The structured-data finding is the robust one. Every number here comes from the saved scan, and any URL can be re-run on the free tool.
Frequently asked questions
Is poor AI visibility just an India problem?
No. Across 24 businesses in the US, Canada and Australia, the sites with no structured data scored a median of 58.5 against 88.5 for the sites that had it, a 30-point gap. The same legibility problem shows up in mature markets.
Which industries scored worst for AI search?
Dental had the lowest median at 70.5 and restaurants 74, while med spas led at 98 and law firms and hotels sat around 87 and 88. Small samples, but healthcare and hospitality sites had the most missing markup.
If half the sites scored an A, is there still a problem?
Yes. Twelve of 24 scored an A, but the spread ran from 29 to 100, and 88% still had no FAQ markup. The businesses that get named by AI are the legible ones.
How was this measured?
Each homepage was fetched once and scored against a fixed list of AI-readability signals with the free Report Card on vishkul.com, on 17 July 2026, no JavaScript executed. The score measures machine-readability, not the quality of the business.